minor changes to JSON schema output

topic_extr.py

@@ -0,0 +1,833 @@
+#!/usr/bin/env python3
+import os
+import re
+import gc
+import json
+import logging
+import fitz
+import boto3
+import base64
+import time
+import asyncio
+import tempfile
+import requests
+from io import BytesIO
+from typing import List, Dict, Any
+
+import torch
+import cv2
+import numpy as np
+
+from google import genai
+from google.genai import types
+
+from magic_pdf.data.dataset import PymuDocDataset
+from magic_pdf.model.doc_analyze_by_custom_model import doc_analyze
+from magic_pdf.data.data_reader_writer.base import DataWriter
+from table_row_extraction import TableExtractor
+
+logging.basicConfig(level=logging.INFO)
+logger = logging.getLogger(__name__)
+logger.setLevel(logging.INFO)
+file_handler = logging.FileHandler("topic_extraction.log")
+file_handler.setFormatter(logging.Formatter("%(asctime)s [%(levelname)s] %(name)s - %(message)s"))
+logger.addHandler(file_handler)
+
+_GEMINI_CLIENT = None
+
+def unify_whitespace(text: str) -> str:
+    return re.sub(r"\s+", " ", text).strip()
+
+def find_all_occurrences(pdf_bytes: bytes, search_text: str) -> List[int]:
+    doc = fitz.open(stream=pdf_bytes, filetype="pdf")
+    st_norm = unify_whitespace(search_text)
+    found = []
+    for i in range(doc.page_count):
+        raw = doc[i].get_text("raw")
+        norm = unify_whitespace(raw)
+        if st_norm in norm:
+            found.append(i)
+    doc.close()
+    return sorted(found)
+
+def create_subset_pdf(original_pdf_bytes: bytes, page_indices: List[int]) -> bytes:
+    if not page_indices:
+        raise ValueError("No page indices provided for subset creation.")
+    doc = fitz.open(stream=original_pdf_bytes, filetype="pdf")
+    new_doc = fitz.open()
+    for p in sorted(set(page_indices)):
+        if 0 <= p < doc.page_count:
+            new_doc.insert_pdf(doc, from_page=p, to_page=p)
+        else:
+            logger.error(f"Page index {p} out of range (0..{doc.page_count - 1}).")
+            raise ValueError(f"Page index {p} out of range.")
+    subset_bytes = new_doc.tobytes()
+    new_doc.close()
+    doc.close()
+    return subset_bytes
+
+class s3Writer:
+    def __init__(self, ak: str, sk: str, bucket: str, endpoint_url: str):
+        self.bucket = bucket
+        self.client = boto3.client(
+            's3',
+            aws_access_key_id=ak,
+            aws_secret_access_key=sk,
+            endpoint_url=endpoint_url
+        )
+
+    def write(self, path: str, data: bytes) -> None:
+        try:
+            file_obj = BytesIO(data)
+            self.client.upload_fileobj(
+                file_obj,
+                self.bucket,
+                path
+            )
+            logger.info(f"Uploaded to S3: {path}")
+        except Exception as e:
+            logger.error(f"Failed to upload to S3: {str(e)}")
+            raise
+
+def preprocess_image(image_data: bytes, max_dim: int = 600, quality: int = 60) -> bytes:
+    arr = np.frombuffer(image_data, np.uint8)
+    img = cv2.imdecode(arr, cv2.IMREAD_COLOR)
+    if img is not None:
+        h, w, _ = img.shape
+        if max(h, w) > max_dim:
+            scale = max_dim / float(max(h, w))
+            new_w = int(w * scale)
+            new_h = int(h * scale)
+            img = cv2.resize(img, (new_w, new_h), interpolation=cv2.INTER_AREA)
+        encode_params = [int(cv2.IMWRITE_JPEG_QUALITY), quality]
+        success, enc = cv2.imencode(".jpg", img, encode_params)
+        if success:
+            return enc.tobytes()
+    return image_data
+
+def call_gemini_for_table_classification(image_data: bytes, api_key: str, max_retries: int = 1) -> str:
+    """
+    Existing Gemini call to classify an image as TWO_COLUMN, THREE_COLUMN, or NO_TABLE.
+    """
+    for attempt in range(max_retries + 1):
+        try:
+            prompt = """You are given an image. Determine if it shows a table that has exactly 2 or 3 columns. 
+The three-column 'table' image includes such key features:
+    - Three columns header
+    - Headers like 'Topics', 'Content', 'Guidelines'
+    - Possibly sections (e.g. 8.4, 9.1)
+The two-column 'table' image includes such key features:
+    - Two columns
+    - Headers like 'Subject content' and 'Additional information'
+    - Possibly sections (e.g. 2.1, 3.4)
+If the image is a relevant table with 2 columns, respond with 'TWO_COLUMN'.
+If the image is a relevant table with 3 columns, respond with 'THREE_COLUMN'.
+If the image does not show a table at all, respond with 'NO_TABLE'.
+Return only one of these exact labels.
+"""
+            global _GEMINI_CLIENT
+            if _GEMINI_CLIENT is None:
+                _GEMINI_CLIENT = genai.Client(api_key=api_key)
+            client = _GEMINI_CLIENT
+
+            resp = client.models.generate_content(
+                model="gemini-2.0-flash",
+                contents=[
+                    {
+                        "parts": [
+                            {"text": prompt},
+                            {
+                                "inline_data": {
+                                    "mime_type": "image/jpeg",
+                                    "data": base64.b64encode(image_data).decode('utf-8')
+                                }
+                            }
+                        ]
+                    }
+                ],
+                config=types.GenerateContentConfig(temperature=0.0)
+            )
+            if resp and resp.text:
+                classification = resp.text.strip().upper()
+                if "THREE" in classification:
+                    return "THREE_COLUMN"
+                elif "TWO" in classification:
+                    return "TWO_COLUMN"
+            return "NO_TABLE"
+        except Exception as e:
+            logger.error(f"Gemini table classification error: {e}")
+            if "503" in str(e):
+                return "NO_TABLE"
+            if attempt < max_retries:
+                time.sleep(0.5)
+            else:
+                return "NO_TABLE"
+
+async def classify_image_async(image_data: bytes, api_key: str, max_retries: int = 1) -> str:
+    loop = asyncio.get_event_loop()
+    preprocessed = preprocess_image(image_data)
+    return await loop.run_in_executor(None, call_gemini_for_table_classification, preprocessed, api_key, max_retries)
+
+
+def call_gemini_for_subtopic_identification(text: str, api_key: str, max_retries: int = 1) -> dict:
+    """
+    Sends the recognized text from a specification table to Gemini,
+    asking it to identify the main topic (like '2 Algebra and functions')
+    and subtopics (like '2.5', '3.4', etc.).
+
+    Returns a dict of the form:
+    {
+      "title": "2 Algebra and functions",
+      "subtopics": ["2.5", "2.6", ...]
+    }
+
+    If Gemini can't find anything, it might return empty strings or lists.
+    """
+    for attempt in range(max_retries + 1):
+        try:
+            prompt = f"""
+You are given text extracted from a table that represents topics and subtopics from an educational curriculum.
+The text may include a main topic heading in the format: "<number> <Topic Name>", for example, "2 Algebra and functions".
+It may also include subtopics in the format of "<number>.<number>", such as "2.5", "3.4", etc.
+Extract and output a valid JSON object with exactly two keys:
+  - "title": the main topic heading (if found). If not found, use an empty string.
+  - "subtopics": an array of strings representing each subtopic number extracted from the text.
+Output exactly in this JSON format with no additional text. For example:
+{
+  "title": "2 Algebra and functions",
+  "subtopics": ["2.5", "2.6"]
+}
+Text:
+{text}
+"""
+            global _GEMINI_CLIENT
+            if _GEMINI_CLIENT is None:
+                _GEMINI_CLIENT = genai.Client(api_key=api_key)
+            client = _GEMINI_CLIENT
+
+            resp = client.models.generate_content(
+                model="gemini-2.0-flash",
+                contents=[prompt],
+                config=types.GenerateContentConfig(temperature=0.0)
+            )
+
+            if not resp or not resp.text:
+                # If Gemini gives no response, fallback to empty.
+                return {"title": "", "subtopics": []}
+
+            raw = resp.text.strip()
+
+            # Attempt to parse raw as JSON
+            try:
+                data = json.loads(raw)
+                # Guarantee the structure we want
+                title = data.get("title", "")
+                subs = data.get("subtopics", [])
+                if not isinstance(subs, list):
+                    subs = []
+                return {"title": title, "subtopics": subs}
+            except Exception:
+                # If JSON parse fails, return empty
+                return {"title": "", "subtopics": []}
+
+        except Exception as e:
+            # If there's an error or a 503, we can retry or bail out
+            if attempt < max_retries:
+                time.sleep(0.5)
+            else:
+                return {"title": "", "subtopics": []}
+
+
+
+class S3ImageWriter(DataWriter):
+    def __init__(self, s3_writer: s3Writer, base_path: str, gemini_api_key: str):
+        self.s3_writer = s3_writer
+        self.base_path = base_path if base_path.endswith("/") else base_path + "/"
+        self.gemini_api_key = gemini_api_key
+        self.descriptions = {}
+        self._img_count = 0
+        self.extracted_tables = {}
+        # New attribute to store final subtopic JSON
+        self.extracted_subtopics = {}
+
+    def write(self, path: str, data: bytes) -> None:
+        self._img_count += 1
+        unique_id = f"img_{self._img_count}.jpg"
+        s3_key = f"{self.base_path}{unique_id}"
+        self.s3_writer.write(s3_key, data)
+        self.descriptions[path] = {
+            "data": data,
+            "s3_path": s3_key,
+            "table_classification": "NO_TABLE",
+            "final_alt": ""
+        }
+
+    async def post_process_async(self, key: str, md_content: str) -> str:
+        logger.info("Classifying images to detect tables.")
+        tasks = {
+            p: asyncio.create_task(classify_image_async(info["data"], self.gemini_api_key))
+            for p, info in self.descriptions.items()
+        }
+        results = await asyncio.gather(*tasks.values(), return_exceptions=True)
+        for p, result in zip(tasks.keys(), results):
+            if isinstance(result, Exception):
+                logger.error(f"Table classification error for {p}: {result}")
+                self.descriptions[p]['table_classification'] = "NO_TABLE"
+            else:
+                self.descriptions[p]['table_classification'] = result
+
+        # 2) Replace the original markdown references with alt text
+        for p, info in self.descriptions.items():
+            cls = info['table_classification']
+            if cls == "TWO_COLUMN":
+                info['final_alt'] = "HAS TO BE PROCESSED - two column table"
+            elif cls == "THREE_COLUMN":
+                info['final_alt'] = "HAS TO BE PROCESSED - three column table"
+            else:
+                info['final_alt'] = "NO_TABLE image"
+            md_content = md_content.replace(f"![]({key}{p})", f"![{info['final_alt']}]({info['s3_path']})")
+
+        md_content = await self._process_table_images_in_markdown(key, md_content)
+
+        # Filter final lines to keep only lines with images
+        final_lines = [
+            line.strip() for line in md_content.split("\n")
+            if re.match(r"^\!\[.*\]\(.*\)", line.strip())
+        ]
+        return "\n".join(final_lines)
+
+
+    async def _process_table_images_in_markdown(self, key: str, md_content: str) -> str:
+        pat = r"!\[HAS TO BE PROCESSED - (two|three) column table\]\(([^)]+)\)"
+        matches = re.findall(pat, md_content, flags=re.IGNORECASE)
+        if not matches:
+            return md_content
+
+        for (col_type, s3_key) in matches:
+            logger.info(f"Processing table image: {s3_key}, columns={col_type}")
+            img_data = None
+            for desc in self.descriptions.values():
+                if desc.get("s3_path") == s3_key:
+                    img_data = desc.get("data")
+                    break
+            if img_data is None:
+                logger.warning(f"No image data found for S3 key {s3_key}. Skipping.")
+                continue
+
+            with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file:
+                temp_file.write(img_data)
+                temp_path = temp_file.name
+
+            try:
+                if col_type.lower() == 'two':
+                    extractor = TableExtractor(
+                        skip_header=True,
+                        merge_two_col_rows=True,
+                        enable_subtopic_merge=True,
+                        subtopic_threshold=0.2
+                    )
+                else:
+                    extractor = TableExtractor(
+                        skip_header=True,
+                        merge_two_col_rows=False,
+                        enable_subtopic_merge=False,
+                        subtopic_threshold=0.2
+                    )
+                row_boxes = extractor.process_image(temp_path)
+
+                snippet = ["**Extracted table cells:**"]
+                cell_texts = []
+                for i, row in enumerate(row_boxes):
+                    for j, box in enumerate(row):
+                        cell_key = f"{self.base_path}cells/table_s3_{os.path.basename(s3_key)}_r{i}_c{j}.jpg"
+                        self.s3_writer.write(cell_key, img_data)  # or cell_data if you truly cropped
+
+                        text = "..."  # placeholder
+                        cell_texts.append(text)
+
+                        snippet.append(f"![Row {i} Col {j}]({cell_key})")
+
+                combined_text = "\n".join(cell_texts)
+
+                subtopic_info = call_gemini_for_subtopic_identification(combined_text, self.gemini_api_key)
+
+                # subtopic_info might be: {"title": "2 Algebra and functions", "subtopics": ["2.5"]}
+                final_json = {
+                    "title": subtopic_info.get("title", ""),
+                    "contents": [
+                        {"type": "image", "key": s3_key}
+                    ],
+                    "children": []
+                }
+                for st in subtopic_info.get("subtopics", []):
+                    final_json["children"].append({
+                        "title": st,
+                        "contents": [
+                            {"type": "image", "key": f"subtopic_{st}_example.jpg"}
+                        ]
+                    })
+
+                self.extracted_subtopics[s3_key] = final_json
+
+                # Replace the original table image line in the markdown with the snippet
+                new_snip = "\n".join(snippet)
+                old_line = f"![HAS TO BE PROCESSED - {col_type} column table]({s3_key})"
+                md_content = md_content.replace(old_line, new_snip)
+
+            except Exception as e:
+                logger.error(f"Error processing table image {s3_key}: {e}")
+            finally:
+                os.remove(temp_path)
+
+        return md_content
+
+    def post_process(self, key: str, md_content: str) -> str:
+        return asyncio.run(self.post_process_async(key, md_content))
+
+
+class LocalImageWriter(DataWriter):
+    def __init__(self, output_folder: str, gemini_api_key: str):
+        self.output_folder = output_folder
+        os.makedirs(self.output_folder, exist_ok=True)
+        self.descriptions = {}
+        self._img_count = 0
+        self.gemini_api_key = gemini_api_key
+
+        self.extracted_tables = {}
+
+    def write(self, path: str, data: bytes) -> None:
+        self._img_count += 1
+        unique_id = f"img_{self._img_count}.jpg"
+        self.descriptions[path] = {
+            "data": data,
+            "relative_path": unique_id,
+            "table_classification": "NO_TABLE",
+            "final_alt": ""
+        }
+        # Also save the original image locally for testing.
+        image_path = os.path.join(self.output_folder, unique_id)
+        with open(image_path, "wb") as f:
+            f.write(data)
+
+    async def post_process_async(self, key: str, md_content: str) -> str:
+        logger.info("Classifying images to detect tables.")
+        tasks = []
+        for p, info in self.descriptions.items():
+            tasks.append((p, classify_image_async(info["data"], self.gemini_api_key)))
+        for p, task in tasks:
+            try:
+                classification = await task
+                self.descriptions[p]['table_classification'] = classification
+            except Exception as e:
+                logger.error(f"Table classification error: {e}")
+                self.descriptions[p]['table_classification'] = "NO_TABLE"
+        for p, info in self.descriptions.items():
+            cls = info['table_classification']
+            if cls == "TWO_COLUMN":
+                info['final_alt'] = "HAS TO BE PROCESSED - two column table"
+            elif cls == "THREE_COLUMN":
+                info['final_alt'] = "HAS TO BE PROCESSED - three column table"
+            else:
+                info['final_alt'] = "NO_TABLE image"
+            md_content = md_content.replace(f"![]({key}{p})", f"![{info['final_alt']}]({info['relative_path']})")
+        md_content = self._process_table_images_in_markdown(md_content)
+        final_lines = []
+        for line in md_content.split("\n"):
+            if re.match(r"^\!\[.*\]\(.*\)", line.strip()):
+                final_lines.append(line.strip())
+        return "\n".join(final_lines)
+
+    def _process_table_images_in_markdown(self, md_content: str) -> str:
+        pat = r"!\[HAS TO BE PROCESSED - (two|three) column table\]\(([^)]+)\)"
+        matches = re.findall(pat, md_content, flags=re.IGNORECASE)
+        if not matches:
+            return md_content
+        for (col_type, image_id) in matches:
+            logger.info(f"Processing table image => {image_id}, columns={col_type}")
+            temp_path = os.path.join(self.output_folder, image_id)
+            desc_item = None
+            for k, val in self.descriptions.items():
+                if val["relative_path"] == image_id:
+                    desc_item = val
+                    break
+            if not desc_item:
+                logger.warning(f"No matching image data for {image_id}, skipping extraction.")
+                continue
+            if not os.path.exists(temp_path):
+                with open(temp_path, "wb") as f:
+                    f.write(desc_item["data"])
+            try:
+                if col_type.lower() == 'two':
+                    extractor = TableExtractor(
+                        skip_header=True,
+                        merge_two_col_rows=True,
+                        enable_subtopic_merge=True,
+                        subtopic_threshold=0.2
+                    )
+                else:
+                    extractor = TableExtractor(
+                        skip_header=True,
+                        merge_two_col_rows=False,
+                        enable_subtopic_merge=False,
+                        subtopic_threshold=0.2
+                    )
+                row_boxes = extractor.process_image(temp_path)
+                out_folder = temp_path + "_rows"
+                os.makedirs(out_folder, exist_ok=True)
+                extractor.save_extracted_cells(temp_path, row_boxes, out_folder)
+                # List all extracted cell images relative to the output folder.
+                extracted_cells = []
+                for root, dirs, files in os.walk(out_folder):
+                    for file in files:
+                        rel_path = os.path.relpath(os.path.join(root, file), self.output_folder)
+                        extracted_cells.append(rel_path)
+                # Save mapping for testing.
+                self.extracted_tables[image_id] = extracted_cells
+                snippet = ["**Extracted table cells:**"]
+                for i, row in enumerate(row_boxes):
+                    row_dir = os.path.join(out_folder, f"row_{i}")
+                    for j, _ in enumerate(row):
+                        cell_file = f"col_{j}.jpg"
+                        cell_path = os.path.join(row_dir, cell_file)
+                        relp = os.path.relpath(cell_path, self.output_folder)
+                        snippet.append(f"![Row {i} Col {j}]({relp})")
+                new_snip = "\n".join(snippet)
+                old_line = f"![HAS TO BE PROCESSED - {col_type} column table]({image_id})"
+                md_content = md_content.replace(old_line, new_snip)
+            except Exception as e:
+                logger.error(f"Error processing table image {image_id}: {e}")
+            finally:
+                if os.path.exists(temp_path):
+                    os.remove(temp_path)
+        return md_content
+
+    def post_process(self, key: str, md_content: str) -> str:
+        return asyncio.run(self.post_process_async(key, md_content))
+
+class GeminiTopicExtractor:
+    def __init__(self, api_key: str = None, num_pages: int = 14):
+        self.api_key = api_key or os.getenv("GEMINI_API_KEY", "")
+        self.num_pages = num_pages
+
+    def extract_subtopics(self, pdf_path: str) -> Dict[str, List[int]]:
+        first_pages_text = self._read_first_pages_raw(pdf_path, self.num_pages)
+        if not first_pages_text.strip():
+            logger.error("No text from first pages => cannot extract subtopics.")
+            return {}
+        prompt = f"""
+You have the first pages of a PDF specification, including a table of contents.
+Instructions:
+1. Identify the 'Contents' section listing all topics, subtopics, and their corresponding pages.
+2. Identify the major academic subtopics (common desired topic names "Paper X", "Theme X", "Content of X", "AS Unit X", "A2 Unit X", or similar headings).
+3. For each subtopic, give the range of pages [start_page, end_page] (1-based) from the table of contents.
+4. Output only valid JSON of the form:
+    {{
+    "Subtopic A": [start_page, end_page],
+    "Subtopic B": [start_page, end_page]
+    }}
+5. If you can't find any subtopics, return an empty JSON.
+Important notes:
+- The correct "end_page" must be the page number of the next topic or subtopic minus 1.
+- The final output must be valid JSON only, with no extra text or code blocks.
+Examples:
+1. Given this table of contents:
+1 Introduction – 2
+    Why choose Edexcel A Level Mathematics? - 2
+    Supporting you in planning and implementing this qualification - 3
+    Qualification at a glance - 5
+2 Subject content and assessment information – 7
+    Paper 1 and Paper 2: Pure Mathematics - 11
+    Paper 3: Statistics and Mechanics - 30
+    Assessment Objectives - 40
+3 Administration and general information – 42
+    Entries - 42
+    Access arrangements, reasonable adjustments, special consideration and malpractice - 42
+    Student recruitment and progression - 45
+Appendix 1: Formulae – 49
+Appendix 2: Notation – 53
+Appendix 3: Use of calculators – 59
+Appendix 4: Assessment Objectives – 60
+Appendix 5: The context for the development of this qualification – 62
+Appendix 6: Transferable skills – 64
+Appendix 7: Level 3 Extended Project qualification – 65
+Appendix 8: Codes – 67
+The correct output should be:
+{{
+    "Paper 1 and Paper 2: Pure Mathematics": [11, 29],
+    "Paper 3: Statistics and Mechanics": [30, 42]
+}}
+2. Given this table of contents:
+Qualification at a glance – 1
+    Assessment Objectives and weightings - 4
+Knowledge, skills and understanding – 5
+    Theme 1: Introduction to markets and market failure - 5
+    Theme 2: The UK economy – performance and policies - 11
+    Theme 3: Business behaviour and the labour market - 21
+    Theme 4: A global perspective - 29
+Assessment – 39
+    Assessment summary - 39
+    Assessment objectives - 41
+    Assessment overview - 42
+    Breakdown of assessment objectives - 42
+        Synoptic assessment - 43
+        Discount code and performance tables - 43
+        Access arrangements, reasonable adjustments and special consideration - 44
+        Malpractice - 45
+        Equality Act 2010 and Pearson equality policy - 45
+        Synoptic assessment - 46
+        Awarding and reporting - 47
+Other information – 49
+    Student recruitment -49
+    Prior learning and other requirements -49
+    Progression - 49
+Appendix 1: Transferable skills – 53
+Appendix 2: Level 3 Extended Project qualification – 55
+Appendix 3: Quantitative skills – 59
+Appendix 4: Codes – 61
+Appendix 5: Index – 63
+The correct output should be:
+{{
+    "Theme 1: Introduction to markets and market failure": [5, 10],
+    "Theme 2: The UK economy – performance and policies": [11, 20],
+    "Theme 3: Business behaviour and the labour market": [21, 28],
+    "Theme 4: A global perspective": [29, 38]
+}}
+3. You might also see sections like:
+2.1 AS Unit 1 11
+2.2 AS Unit 2 18
+2.3 A2 Unit 3 24
+2.4 A2 Unit 4 31
+In that scenario, your output might look like:
+{{
+    "2.1 AS Unit 1": [11, 17],
+    "2.2 AS Unit 2": [18, 23],
+    "2.3 A2 Unit 3": [24, 30],
+    "2.4 A2 Unit 4": [31, 35]
+}}
+4. Another example might list subtopics:
+3.1 Overarching themes 11
+3.2 A: Proof 12
+3.3 B: Algebra and functions 13
+3.4 C: Coordinate geometry in the ( x , y ) plane 14
+3.5 D: Sequences and series 15
+3.6 E: Trigonometry 16
+3.7 F: Exponentials and logarithms 17
+3.8 G: Differentiation 18
+3.9 H: Integration 19
+3.10 I: Numerical methods 20
+3.11 J: Vectors 20
+3.12 K: Statistical sampling 21
+3.13 L: Data presentation and interpretation 21
+3.14 M: Probability 22
+3.15 N: Statistical distributions 23
+3.16 O: Statistical hypothesis testing 23
+3.17 P: Quantities and units in mechanics 24
+3.18 Q: Kinematics 24
+3.19 R: Forces and Newton’s laws 24
+3.20 S: Moments 25
+3.21 Use of data in statistics 26
+Here the correct output might look like:
+{{
+    "A: Proof": [12, 12],
+    "B: Algebra and functions": [13, 13],
+    ...
+}}
+Now, extract topics from this text:
+{first_pages_text}
+"""
+        global _GEMINI_CLIENT
+        if _GEMINI_CLIENT is None:
+            _GEMINI_CLIENT = genai.Client(api_key=self.api_key)
+        client = _GEMINI_CLIENT
+        try:
+            response = client.models.generate_content(
+                model="gemini-2.0-flash",
+                contents=[prompt],
+                config=types.GenerateContentConfig(temperature=0.0)
+            )
+            if not response or not response.text:
+                logger.warning("No text from LLM => returning empty subtopics.")
+                return {}
+            raw_json = response.text.strip()
+            cleaned = raw_json.replace("```json", "").replace("```", "")
+            try:
+                data = json.loads(cleaned)
+            except Exception as json_err:
+                logger.error(f"JSON parsing error: {json_err}")
+                return {}
+            final_dict = {}
+            found_sub_dict = None
+            for k, v in data.items():
+                if isinstance(v, dict):
+                    found_sub_dict = v
+                    break
+            if found_sub_dict is not None:
+                for subk, rng in found_sub_dict.items():
+                    if isinstance(rng, list) and len(rng) == 2:
+                        final_dict[subk] = rng
+            else:
+                for subk, rng in data.items():
+                    if isinstance(rng, list) and len(rng) == 2:
+                        final_dict[subk] = rng
+            return final_dict
+        except Exception as e:
+            logger.error(f"Gemini subtopic extraction error: {e}")
+            return {}
+
+    def _read_first_pages_raw(self, pdf_path: str, num_pages: int) -> str:
+        text_parts = []
+        try:
+            if pdf_path.startswith("http://") or pdf_path.startswith("https://"):
+                response = requests.get(pdf_path)
+                if response.status_code != 200:
+                    logger.error("Failed to download PDF from %s. Status code: %d", pdf_path, response.status_code)
+                    return ""
+                pdf_bytes = response.content
+            else:
+                with open(pdf_path, "rb") as f:
+                    pdf_bytes = f.read()
+            doc = fitz.open(stream=pdf_bytes, filetype="pdf")
+            pages_to_read = min(num_pages, doc.page_count)
+            for i in range(pages_to_read):
+                raw_text = doc[i].get_text("raw")
+                text_parts.append(raw_text)
+            doc.close()
+        except Exception as e:
+            logger.error(f"Could not open PDF: {e}")
+        return "\n".join(text_parts)
+
+
+class MineruNoTextProcessor:
+    def __init__(self, output_folder: str, gemini_api_key: str):
+        self.output_folder = output_folder
+        os.makedirs(self.output_folder, exist_ok=True)
+        self.layout_model = "doclayout_yolo"
+        self.formula_enable = True
+        self.table_enable = False
+        self.language = "en"
+
+        self.subtopic_extractor = GeminiTopicExtractor(api_key=gemini_api_key, num_pages=20)
+        self.gemini_api_key = gemini_api_key or os.getenv("GEMINI_API_KEY", "")
+
+        self.use_s3 = True
+        self.s3_writer = s3Writer(
+            ak=os.getenv("S3_ACCESS_KEY"),
+            sk=os.getenv("S3_SECRET_KEY"),
+            bucket="quextro-resources",
+            endpoint_url=os.getenv("S3_ENDPOINT")
+        )
+
+    def cleanup_gpu(self):
+        try:
+            gc.collect()
+            torch.cuda.empty_cache()
+            logger.info("GPU memory cleaned up.")
+        except Exception as e:
+            logger.error(f"Error during GPU cleanup: {e}")
+
+    def process(self, pdf_path: str) -> Dict[str, Any]:
+        logger.info(f"Processing PDF: {pdf_path}")
+        try:
+            # 1) Possibly call subtopic_extractor on first pages to find subtopics in the PDF as a whole
+            subtopics = self.subtopic_extractor.extract_subtopics(pdf_path)
+            logger.info(f"Gemini returned subtopics: {subtopics}")
+
+            if pdf_path.startswith("http://") or pdf_path.startswith("https://"):
+                response = requests.get(pdf_path)
+                if response.status_code != 200:
+                    logger.error("Failed to download PDF from %s. Status code: %d", pdf_path, response.status_code)
+                    raise Exception(f"Failed to download PDF: {pdf_path}")
+                pdf_bytes = response.content
+                logger.info("Downloaded %d bytes for pdf_url='%s'", len(pdf_bytes), pdf_path)
+            else:
+                with open(pdf_path, "rb") as f:
+                    pdf_bytes = f.read()
+                logger.info("Loaded %d bytes from local file '%s'", len(pdf_bytes), pdf_path)
+
+            doc = fitz.open(stream=pdf_bytes, filetype="pdf")
+            total_pages = doc.page_count
+            doc.close()
+
+            # 3) Decide which pages to process
+            final_pages = set()
+            if not subtopics:
+                # fallback
+                final_pages = set(range(total_pages))
+            else:
+                offset_candidates = []
+                for subname, rng in subtopics.items():
+                    start_p, _ = rng
+                    occs = find_all_occurrences(pdf_bytes, subname)
+                    for p in occs:
+                        candidate = p - (start_p - 1)
+                        if candidate > 0:
+                            offset_candidates.append(candidate)
+                if offset_candidates:
+                    try:
+                        from statistics import mode
+                        global_offset = mode(offset_candidates)
+                    except:
+                        from statistics import median
+                        global_offset = int(median(offset_candidates))
+                else:
+                    global_offset = 0
+
+                logger.info(f"Computed global offset: {global_offset}")
+                for subname, rng in subtopics.items():
+                    if not (isinstance(rng, list) and len(rng) == 2):
+                        continue
+                    start_p, end_p = rng
+                    if start_p > end_p:
+                        continue
+                    s0 = (start_p - 1) + global_offset
+                    e0 = (end_p - 1) + global_offset
+                    for pp in range(s0, e0 + 1):
+                        final_pages.add(pp)
+
+            if not final_pages:
+                final_pages = set(range(total_pages))
+
+            logger.info(f"Processing pages (0-based): {sorted(final_pages)}")
+            subset_pdf_bytes = create_subset_pdf(pdf_bytes, sorted(final_pages))
+
+            # 4) Analyze and produce markdown
+            dataset = PymuDocDataset(subset_pdf_bytes)
+            inference = doc_analyze(
+                dataset,
+                ocr=True,
+                lang=self.language,
+                layout_model=self.layout_model,
+                formula_enable=self.formula_enable,
+                table_enable=self.table_enable
+            )
+            writer = S3ImageWriter(self.s3_writer, "/topic-extraction", self.gemini_api_key)
+            md_prefix = "/topic-extraction/"
+            pipe_result = inference.pipe_ocr_mode(writer, lang=self.language)
+            md_content = pipe_result.get_markdown(md_prefix)
+            final_markdown = writer.post_process(md_prefix, md_content)
+
+            subtopic_list = list(writer.extracted_subtopics.values())
+
+            out_path = os.path.join(self.output_folder, "final_subtopics.json")
+            with open(out_path, "w", encoding="utf-8") as f:
+                json.dump(subtopic_list, f, indent=2)
+            logger.info(f"Final subtopics JSON saved locally at {out_path}")
+
+            return {
+                "final_markdown": final_markdown,
+                "subtopics_extracted": subtopic_list
+            }
+        finally:
+            self.cleanup_gpu()
+
+if __name__ == "__main__":
+    input_pdf = "/home/user/app/input_output/a-level-pearson-mathematics-specification.pdf"
+    output_dir = "/home/user/app/we"
+    gemini_key = os.getenv("GEMINI_API_KEY", "AIzaSyDtoakpXa2pjJwcQB6TJ5QaXHNSA5JxcrU")
+    try:
+        processor = MineruNoTextProcessor(output_folder=output_dir, gemini_api_key=gemini_key)
+        result = processor.process(input_pdf)
+        logger.info("Processing completed successfully.")
+        # The result includes final_markdown and subtopics_extracted
+    except Exception as e:
+        logger.error(f"Processing failed: {e}")

topic_extraction.log

@@ -728,3 +728,467 @@ and series'. Using page 7.
 2025-03-02 15:59:11,730 [WARNING] __main__ - No suitable occurrence for '2.4 A2 Unit 4'. Using page 3.
 2025-03-02 15:59:11,730 [INFO] __main__ - Processing pages (0-based): [10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34]
 2025-03-02 15:59:48,050 [INFO] __main__ - GPU memory cleaned up.
+2025-03-03 13:26:57,282 [INFO] __main__ - Processing PDF: /home/user/app/input_output/a-level-pearson-mathematics-specification.pdf
+2025-03-03 13:26:58,095 [INFO] __main__ - Gemini returned subtopics: {'Paper 1 and Paper 2: Pure Mathematics': [11, 29], 'Paper 3: Statistics and Mechanics': [30, 40]}
+2025-03-03 13:26:58,096 [INFO] __main__ - Loaded 1135473 bytes from local file '/home/user/app/input_output/a-level-pearson-mathematics-specification.pdf'
+2025-03-03 13:26:58,447 [INFO] __main__ - Computed global offset: 4
+2025-03-03 13:26:58,447 [INFO] __main__ - Processing pages (0-based): [14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43]
+2025-03-03 13:28:51,814 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_1.jpg
+2025-03-03 13:28:53,673 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_2.jpg
+2025-03-03 13:28:54,260 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_3.jpg
+2025-03-03 13:28:54,783 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_4.jpg
+2025-03-03 13:28:55,347 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_5.jpg
+2025-03-03 13:28:55,931 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_6.jpg
+2025-03-03 13:28:56,425 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_7.jpg
+2025-03-03 13:28:57,012 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_8.jpg
+2025-03-03 13:28:57,496 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_9.jpg
+2025-03-03 13:28:58,025 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_10.jpg
+2025-03-03 13:28:58,574 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_11.jpg
+2025-03-03 13:28:59,112 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_12.jpg
+2025-03-03 13:28:59,574 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_13.jpg
+2025-03-03 13:29:00,155 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_14.jpg
+2025-03-03 13:29:00,690 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_15.jpg
+2025-03-03 13:29:01,291 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_16.jpg
+2025-03-03 13:29:01,867 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_17.jpg
+2025-03-03 13:29:02,177 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_18.jpg
+2025-03-03 13:29:02,676 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_19.jpg
+2025-03-03 13:29:03,274 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_20.jpg
+2025-03-03 13:29:03,849 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_21.jpg
+2025-03-03 13:29:04,424 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_22.jpg
+2025-03-03 13:29:04,883 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_23.jpg
+2025-03-03 13:29:05,416 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_24.jpg
+2025-03-03 13:29:05,965 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_25.jpg
+2025-03-03 13:29:06,408 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_26.jpg
+2025-03-03 13:29:06,899 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_27.jpg
+2025-03-03 13:29:07,492 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_28.jpg
+2025-03-03 13:29:07,914 [INFO] __main__ - Classifying images to detect tables.
+2025-03-03 13:29:11,945 [INFO] __main__ - Processing table image: /topic-extraction/img_1.jpg, columns=three
+2025-03-03 13:29:15,280 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r0_c0.jpg
+2025-03-03 13:29:15,569 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r0_c1.jpg
+2025-03-03 13:29:15,771 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r1_c0.jpg
+2025-03-03 13:29:16,000 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r1_c1.jpg
+2025-03-03 13:29:16,001 [ERROR] __main__ - Error processing table image /topic-extraction/img_1.jpg: No module named 'your_module'
+2025-03-03 13:29:16,002 [INFO] __main__ - Processing table image: /topic-extraction/img_2.jpg, columns=three
+2025-03-03 13:29:19,539 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r0_c0.jpg
+2025-03-03 13:29:19,840 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r0_c1.jpg
+2025-03-03 13:29:20,032 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r1_c0.jpg
+2025-03-03 13:29:20,254 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r2_c0.jpg
+2025-03-03 13:29:20,493 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r3_c0.jpg
+2025-03-03 13:29:20,493 [ERROR] __main__ - Error processing table image /topic-extraction/img_2.jpg: No module named 'your_module'
+2025-03-03 13:29:20,494 [INFO] __main__ - Processing table image: /topic-extraction/img_3.jpg, columns=three
+2025-03-03 13:29:23,853 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_3.jpg_r0_c0.jpg
+2025-03-03 13:29:24,160 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_3.jpg_r0_c1.jpg
+2025-03-03 13:29:24,408 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_3.jpg_r1_c0.jpg
+2025-03-03 13:29:24,409 [ERROR] __main__ - Error processing table image /topic-extraction/img_3.jpg: No module named 'your_module'
+2025-03-03 13:29:24,410 [INFO] __main__ - Processing table image: /topic-extraction/img_4.jpg, columns=three
+2025-03-03 13:29:27,500 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r0_c0.jpg
+2025-03-03 13:29:27,818 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r0_c1.jpg
+2025-03-03 13:29:28,045 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r1_c0.jpg
+2025-03-03 13:29:28,266 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r1_c1.jpg
+2025-03-03 13:29:28,268 [ERROR] __main__ - Error processing table image /topic-extraction/img_4.jpg: No module named 'your_module'
+2025-03-03 13:29:28,268 [INFO] __main__ - Processing table image: /topic-extraction/img_5.jpg, columns=three
+2025-03-03 13:29:31,772 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r0_c0.jpg
+2025-03-03 13:29:32,031 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r0_c1.jpg
+2025-03-03 13:29:32,220 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r1_c0.jpg
+2025-03-03 13:29:32,432 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r1_c1.jpg
+2025-03-03 13:29:32,646 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r2_c0.jpg
+2025-03-03 13:29:32,648 [ERROR] __main__ - Error processing table image /topic-extraction/img_5.jpg: No module named 'your_module'
+2025-03-03 13:29:32,648 [INFO] __main__ - Processing table image: /topic-extraction/img_6.jpg, columns=three
+2025-03-03 13:29:36,400 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r0_c0.jpg
+2025-03-03 13:29:36,689 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r0_c1.jpg
+2025-03-03 13:29:36,923 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r1_c0.jpg
+2025-03-03 13:29:37,135 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r1_c1.jpg
+2025-03-03 13:29:37,137 [ERROR] __main__ - Error processing table image /topic-extraction/img_6.jpg: No module named 'your_module'
+2025-03-03 13:29:37,137 [INFO] __main__ - Processing table image: /topic-extraction/img_7.jpg, columns=three
+2025-03-03 13:29:40,804 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r0_c0.jpg
+2025-03-03 13:29:41,126 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r0_c1.jpg
+2025-03-03 13:29:41,336 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r1_c0.jpg
+2025-03-03 13:29:41,527 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r2_c0.jpg
+2025-03-03 13:29:41,750 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r2_c1.jpg
+2025-03-03 13:29:41,752 [ERROR] __main__ - Error processing table image /topic-extraction/img_7.jpg: No module named 'your_module'
+2025-03-03 13:29:41,752 [INFO] __main__ - Processing table image: /topic-extraction/img_8.jpg, columns=three
+2025-03-03 13:29:45,625 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r0_c0.jpg
+2025-03-03 13:29:45,929 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r0_c1.jpg
+2025-03-03 13:29:46,120 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r0_c2.jpg
+2025-03-03 13:29:46,332 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r1_c0.jpg
+2025-03-03 13:29:46,554 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r1_c1.jpg
+2025-03-03 13:29:46,785 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r1_c2.jpg
+2025-03-03 13:29:46,973 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r2_c0.jpg
+2025-03-03 13:29:47,161 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r2_c1.jpg
+2025-03-03 13:29:47,371 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r3_c0.jpg
+2025-03-03 13:29:47,604 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r3_c1.jpg
+2025-03-03 13:29:47,860 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r4_c0.jpg
+2025-03-03 13:29:48,073 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r4_c1.jpg
+2025-03-03 13:29:48,293 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r5_c0.jpg
+2025-03-03 13:29:48,527 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r5_c1.jpg
+2025-03-03 13:29:48,528 [ERROR] __main__ - Error processing table image /topic-extraction/img_8.jpg: No module named 'your_module'
+2025-03-03 13:29:48,529 [INFO] __main__ - Processing table image: /topic-extraction/img_9.jpg, columns=three
+2025-03-03 13:29:52,131 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r0_c0.jpg
+2025-03-03 13:29:52,394 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r0_c1.jpg
+2025-03-03 13:29:52,608 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r0_c2.jpg
+2025-03-03 13:29:52,828 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r1_c0.jpg
+2025-03-03 13:29:53,040 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r1_c1.jpg
+2025-03-03 13:29:53,230 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r2_c0.jpg
+2025-03-03 13:29:53,483 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r2_c1.jpg
+2025-03-03 13:29:53,712 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r3_c0.jpg
+2025-03-03 13:29:53,907 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r3_c1.jpg
+2025-03-03 13:29:53,908 [ERROR] __main__ - Error processing table image /topic-extraction/img_9.jpg: No module named 'your_module'
+2025-03-03 13:29:53,908 [INFO] __main__ - Processing table image: /topic-extraction/img_10.jpg, columns=three
+2025-03-03 13:29:57,203 [INFO] __main__ - GPU memory cleaned up.
+2025-03-03 13:33:11,287 [INFO] __main__ - Processing PDF: /home/user/app/input_output/a-level-pearson-mathematics-specification.pdf
+2025-03-03 13:33:12,086 [INFO] __main__ - Gemini returned subtopics: {'Paper 1 and Paper 2: Pure Mathematics': [11, 29], 'Paper 3: Statistics and Mechanics': [30, 40]}
+2025-03-03 13:33:12,087 [INFO] __main__ - Loaded 1135473 bytes from local file '/home/user/app/input_output/a-level-pearson-mathematics-specification.pdf'
+2025-03-03 13:33:12,450 [INFO] __main__ - Computed global offset: 4
+2025-03-03 13:33:12,450 [INFO] __main__ - Processing pages (0-based): [14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43]
+2025-03-03 13:34:08,590 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_1.jpg
+2025-03-03 13:34:10,443 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_2.jpg
+2025-03-03 13:34:11,003 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_3.jpg
+2025-03-03 13:34:11,551 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_4.jpg
+2025-03-03 13:34:12,104 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_5.jpg
+2025-03-03 13:34:12,646 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_6.jpg
+2025-03-03 13:34:13,132 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_7.jpg
+2025-03-03 13:34:13,599 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_8.jpg
+2025-03-03 13:34:14,153 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_9.jpg
+2025-03-03 13:34:14,847 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_10.jpg
+2025-03-03 13:34:15,304 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_11.jpg
+2025-03-03 13:34:15,831 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_12.jpg
+2025-03-03 13:34:16,257 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_13.jpg
+2025-03-03 13:34:16,888 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_14.jpg
+2025-03-03 13:34:17,395 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_15.jpg
+2025-03-03 13:34:17,984 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_16.jpg
+2025-03-03 13:34:18,529 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_17.jpg
+2025-03-03 13:34:18,824 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_18.jpg
+2025-03-03 13:34:19,296 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_19.jpg
+2025-03-03 13:34:19,924 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_20.jpg
+2025-03-03 13:34:20,496 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_21.jpg
+2025-03-03 13:34:21,104 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_22.jpg
+2025-03-03 13:34:21,531 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_23.jpg
+2025-03-03 13:34:22,061 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_24.jpg
+2025-03-03 13:34:22,624 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_25.jpg
+2025-03-03 13:34:23,058 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_26.jpg
+2025-03-03 13:34:23,603 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_27.jpg
+2025-03-03 13:34:24,201 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_28.jpg
+2025-03-03 13:34:24,606 [INFO] __main__ - Classifying images to detect tables.
+2025-03-03 13:34:28,395 [INFO] __main__ - Processing table image: /topic-extraction/img_1.jpg, columns=three
+2025-03-03 13:34:31,655 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r0_c0.jpg
+2025-03-03 13:34:31,943 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r0_c1.jpg
+2025-03-03 13:34:32,144 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r1_c0.jpg
+2025-03-03 13:34:32,376 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r1_c1.jpg
+2025-03-03 13:34:32,801 [INFO] __main__ - Processing table image: /topic-extraction/img_2.jpg, columns=three
+2025-03-03 13:34:36,018 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r0_c0.jpg
+2025-03-03 13:34:36,281 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r0_c1.jpg
+2025-03-03 13:34:36,504 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r1_c0.jpg
+2025-03-03 13:34:36,725 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r2_c0.jpg
+2025-03-03 13:34:36,916 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r3_c0.jpg
+2025-03-03 13:34:37,320 [INFO] __main__ - Processing table image: /topic-extraction/img_3.jpg, columns=three
+2025-03-03 13:34:40,825 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_3.jpg_r0_c0.jpg
+2025-03-03 13:34:41,120 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_3.jpg_r0_c1.jpg
+2025-03-03 13:34:41,332 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_3.jpg_r1_c0.jpg
+2025-03-03 13:34:41,846 [INFO] __main__ - Processing table image: /topic-extraction/img_4.jpg, columns=three
+2025-03-03 13:34:44,953 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r0_c0.jpg
+2025-03-03 13:34:45,248 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r0_c1.jpg
+2025-03-03 13:34:45,467 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r1_c0.jpg
+2025-03-03 13:34:45,695 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r1_c1.jpg
+2025-03-03 13:34:46,080 [INFO] __main__ - Processing table image: /topic-extraction/img_5.jpg, columns=three
+2025-03-03 13:34:49,588 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r0_c0.jpg
+2025-03-03 13:34:49,913 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r0_c1.jpg
+2025-03-03 13:34:50,102 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r1_c0.jpg
+2025-03-03 13:34:50,361 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r1_c1.jpg
+2025-03-03 13:34:50,608 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r2_c0.jpg
+2025-03-03 13:34:51,003 [INFO] __main__ - Processing table image: /topic-extraction/img_6.jpg, columns=three
+2025-03-03 13:34:54,650 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r0_c0.jpg
+2025-03-03 13:34:54,948 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r0_c1.jpg
+2025-03-03 13:34:55,165 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r1_c0.jpg
+2025-03-03 13:34:55,384 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r1_c1.jpg
+2025-03-03 13:34:55,811 [INFO] __main__ - Processing table image: /topic-extraction/img_7.jpg, columns=three
+2025-03-03 13:34:59,337 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r0_c0.jpg
+2025-03-03 13:34:59,623 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r0_c1.jpg
+2025-03-03 13:34:59,864 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r1_c0.jpg
+2025-03-03 13:35:00,087 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r2_c0.jpg
+2025-03-03 13:35:00,301 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r2_c1.jpg
+2025-03-03 13:35:00,734 [INFO] __main__ - Processing table image: /topic-extraction/img_8.jpg, columns=three
+2025-03-03 13:35:04,169 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r0_c0.jpg
+2025-03-03 13:35:04,466 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r0_c1.jpg
+2025-03-03 13:35:04,693 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r0_c2.jpg
+2025-03-03 13:35:04,880 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r1_c0.jpg
+2025-03-03 13:35:05,111 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r1_c1.jpg
+2025-03-03 13:35:05,414 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r1_c2.jpg
+2025-03-03 13:35:05,607 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r2_c0.jpg
+2025-03-03 13:35:05,841 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r2_c1.jpg
+2025-03-03 13:35:06,062 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r3_c0.jpg
+2025-03-03 13:35:06,273 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r3_c1.jpg
+2025-03-03 13:35:06,497 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r4_c0.jpg
+2025-03-03 13:35:06,737 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r4_c1.jpg
+2025-03-03 13:35:06,967 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r5_c0.jpg
+2025-03-03 13:35:07,191 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r5_c1.jpg
+2025-03-03 13:35:07,636 [INFO] __main__ - Processing table image: /topic-extraction/img_9.jpg, columns=three
+2025-03-03 13:35:11,212 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r0_c0.jpg
+2025-03-03 13:35:11,510 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r0_c1.jpg
+2025-03-03 13:35:11,732 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r0_c2.jpg
+2025-03-03 13:35:11,922 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r1_c0.jpg
+2025-03-03 13:35:12,139 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r1_c1.jpg
+2025-03-03 13:35:12,349 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r2_c0.jpg
+2025-03-03 13:35:12,596 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r2_c1.jpg
+2025-03-03 13:35:12,802 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r3_c0.jpg
+2025-03-03 13:35:13,038 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r3_c1.jpg
+2025-03-03 13:35:13,489 [INFO] __main__ - Processing table image: /topic-extraction/img_10.jpg, columns=three
+2025-03-03 13:35:16,908 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_10.jpg_r0_c0.jpg
+2025-03-03 13:35:17,202 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_10.jpg_r0_c1.jpg
+2025-03-03 13:35:17,420 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_10.jpg_r1_c0.jpg
+2025-03-03 13:35:17,635 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_10.jpg_r2_c0.jpg
+2025-03-03 13:35:17,860 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_10.jpg_r3_c0.jpg
+2025-03-03 13:35:18,244 [INFO] __main__ - Processing table image: /topic-extraction/img_11.jpg, columns=two
+2025-03-03 13:35:21,413 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_11.jpg_r0_c0.jpg
+2025-03-03 13:35:21,722 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_11.jpg_r1_c0.jpg
+2025-03-03 13:35:21,947 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_11.jpg_r2_c0.jpg
+2025-03-03 13:35:22,173 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_11.jpg_r3_c0.jpg
+2025-03-03 13:35:22,430 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_11.jpg_r4_c0.jpg
+2025-03-03 13:35:22,616 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_11.jpg_r5_c0.jpg
+2025-03-03 13:35:23,009 [INFO] __main__ - Processing table image: /topic-extraction/img_12.jpg, columns=three
+2025-03-03 13:35:26,169 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_12.jpg_r0_c0.jpg
+2025-03-03 13:35:26,452 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_12.jpg_r0_c1.jpg
+2025-03-03 13:35:26,674 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_12.jpg_r1_c0.jpg
+2025-03-03 13:35:26,894 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_12.jpg_r1_c1.jpg
+2025-03-03 13:35:27,104 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_12.jpg_r2_c0.jpg
+2025-03-03 13:35:27,342 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_12.jpg_r2_c1.jpg
+2025-03-03 13:35:27,736 [INFO] __main__ - Processing table image: /topic-extraction/img_13.jpg, columns=three
+2025-03-03 13:35:30,898 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_13.jpg_r0_c0.jpg
+2025-03-03 13:35:31,165 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_13.jpg_r0_c1.jpg
+2025-03-03 13:35:31,398 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_13.jpg_r1_c0.jpg
+2025-03-03 13:35:31,601 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_13.jpg_r1_c1.jpg
+2025-03-03 13:35:31,789 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_13.jpg_r2_c0.jpg
+2025-03-03 13:35:32,007 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_13.jpg_r3_c0.jpg
+2025-03-03 13:35:32,449 [INFO] __main__ - Processing table image: /topic-extraction/img_14.jpg, columns=three
+2025-03-03 13:35:36,914 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r0_c0.jpg
+2025-03-03 13:35:37,204 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r0_c1.jpg
+2025-03-03 13:35:37,427 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r1_c0.jpg
+2025-03-03 13:35:37,643 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r1_c1.jpg
+2025-03-03 13:35:37,861 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r2_c0.jpg
+2025-03-03 13:35:38,087 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r3_c0.jpg
+2025-03-03 13:35:38,312 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r4_c0.jpg
+2025-03-03 13:35:38,536 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r4_c1.jpg
+2025-03-03 13:35:38,735 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r5_c0.jpg
+2025-03-03 13:35:39,132 [INFO] __main__ - Processing table image: /topic-extraction/img_15.jpg, columns=three
+2025-03-03 13:35:42,577 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r0_c0.jpg
+2025-03-03 13:35:42,882 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r0_c1.jpg
+2025-03-03 13:35:43,086 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r1_c0.jpg
+2025-03-03 13:35:43,310 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r1_c1.jpg
+2025-03-03 13:35:43,529 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r2_c0.jpg
+2025-03-03 13:35:43,718 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r3_c0.jpg
+2025-03-03 13:35:43,914 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r4_c0.jpg
+2025-03-03 13:35:44,304 [INFO] __main__ - Processing table image: /topic-extraction/img_16.jpg, columns=three
+2025-03-03 13:35:48,617 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r0_c0.jpg
+2025-03-03 13:35:48,915 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r0_c1.jpg
+2025-03-03 13:35:49,124 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r1_c0.jpg
+2025-03-03 13:35:49,350 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r1_c1.jpg
+2025-03-03 13:35:49,578 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r2_c0.jpg
+2025-03-03 13:35:49,801 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r3_c0.jpg
+2025-03-03 13:35:50,013 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r3_c1.jpg
+2025-03-03 13:35:50,239 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r4_c0.jpg
+2025-03-03 13:35:51,085 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r5_c0.jpg
+2025-03-03 13:35:51,516 [INFO] __main__ - Processing table image: /topic-extraction/img_17.jpg, columns=three
+2025-03-03 13:35:55,302 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r0_c0.jpg
+2025-03-03 13:35:55,608 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r0_c1.jpg
+2025-03-03 13:35:55,831 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r1_c0.jpg
+2025-03-03 13:35:56,034 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r2_c0.jpg
+2025-03-03 13:35:56,259 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r2_c1.jpg
+2025-03-03 13:35:56,467 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r3_c0.jpg
+2025-03-03 13:35:56,661 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r4_c0.jpg
+2025-03-03 13:35:56,882 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r5_c0.jpg
+2025-03-03 13:35:57,282 [INFO] __main__ - Processing table image: /topic-extraction/img_18.jpg, columns=three
+2025-03-03 13:35:58,547 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_18.jpg_r0_c0.jpg
+2025-03-03 13:35:58,740 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_18.jpg_r0_c1.jpg
+2025-03-03 13:35:58,934 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_18.jpg_r1_c0.jpg
+2025-03-03 13:35:59,124 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_18.jpg_r1_c1.jpg
+2025-03-03 13:35:59,543 [INFO] __main__ - Processing table image: /topic-extraction/img_19.jpg, columns=three
+2025-03-03 13:36:02,472 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_19.jpg_r0_c0.jpg
+2025-03-03 13:36:02,792 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_19.jpg_r0_c1.jpg
+2025-03-03 13:36:03,024 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_19.jpg_r1_c0.jpg
+2025-03-03 13:36:03,265 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_19.jpg_r1_c1.jpg
+2025-03-03 13:36:03,521 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_19.jpg_r2_c0.jpg
+2025-03-03 13:36:03,745 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_19.jpg_r2_c1.jpg
+2025-03-03 13:36:04,162 [INFO] __main__ - Processing table image: /topic-extraction/img_20.jpg, columns=three
+2025-03-03 13:36:09,253 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_20.jpg_r0_c0.jpg
+2025-03-03 13:36:09,551 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_20.jpg_r0_c1.jpg
+2025-03-03 13:36:09,791 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_20.jpg_r1_c0.jpg
+2025-03-03 13:36:09,983 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_20.jpg_r1_c1.jpg
+2025-03-03 13:36:10,370 [INFO] __main__ - Processing table image: /topic-extraction/img_21.jpg, columns=three
+2025-03-03 13:36:14,808 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_21.jpg_r0_c0.jpg
+2025-03-03 13:36:15,103 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_21.jpg_r0_c1.jpg
+2025-03-03 13:36:15,329 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_21.jpg_r1_c0.jpg
+2025-03-03 13:36:15,570 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_21.jpg_r1_c1.jpg
+2025-03-03 13:36:16,020 [INFO] __main__ - Processing table image: /topic-extraction/img_22.jpg, columns=three
+2025-03-03 13:36:20,786 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r0_c0.jpg
+2025-03-03 13:36:21,053 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r0_c1.jpg
+2025-03-03 13:36:21,295 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r1_c0.jpg
+2025-03-03 13:36:21,539 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r1_c1.jpg
+2025-03-03 13:36:21,753 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r2_c0.jpg
+2025-03-03 13:36:21,966 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r2_c1.jpg
+2025-03-03 13:36:22,203 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r3_c0.jpg
+2025-03-03 13:36:22,623 [INFO] __main__ - Processing table image: /topic-extraction/img_23.jpg, columns=three
+2025-03-03 13:36:25,781 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_23.jpg_r0_c0.jpg
+2025-03-03 13:36:26,112 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_23.jpg_r0_c1.jpg
+2025-03-03 13:36:26,300 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_23.jpg_r1_c0.jpg
+2025-03-03 13:36:26,540 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_23.jpg_r1_c1.jpg
+2025-03-03 13:36:26,953 [INFO] __main__ - Processing table image: /topic-extraction/img_24.jpg, columns=three
+2025-03-03 13:36:30,871 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_24.jpg_r0_c0.jpg
+2025-03-03 13:36:31,169 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_24.jpg_r0_c1.jpg
+2025-03-03 13:36:31,360 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_24.jpg_r1_c0.jpg
+2025-03-03 13:36:31,586 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_24.jpg_r1_c1.jpg
+2025-03-03 13:36:31,839 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_24.jpg_r2_c0.jpg
+2025-03-03 13:36:32,233 [INFO] __main__ - Processing table image: /topic-extraction/img_25.jpg, columns=two
+2025-03-03 13:36:35,440 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_25.jpg_r0_c0.jpg
+2025-03-03 13:36:35,737 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_25.jpg_r1_c0.jpg
+2025-03-03 13:36:35,958 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_25.jpg_r2_c0.jpg
+2025-03-03 13:36:36,186 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_25.jpg_r3_c0.jpg
+2025-03-03 13:36:36,417 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_25.jpg_r4_c0.jpg
+2025-03-03 13:36:36,615 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_25.jpg_r5_c0.jpg
+2025-03-03 13:36:37,031 [INFO] __main__ - Processing table image: /topic-extraction/img_26.jpg, columns=three
+2025-03-03 13:36:40,174 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_26.jpg_r0_c0.jpg
+2025-03-03 13:36:40,461 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_26.jpg_r0_c1.jpg
+2025-03-03 13:36:40,681 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_26.jpg_r1_c0.jpg
+2025-03-03 13:36:40,872 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_26.jpg_r2_c0.jpg
+2025-03-03 13:36:41,298 [INFO] __main__ - Processing table image: /topic-extraction/img_27.jpg, columns=three
+2025-03-03 13:36:44,242 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r0_c0.jpg
+2025-03-03 13:36:44,546 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r0_c1.jpg
+2025-03-03 13:36:44,735 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r1_c0.jpg
+2025-03-03 13:36:44,958 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r1_c1.jpg
+2025-03-03 13:36:45,189 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r2_c0.jpg
+2025-03-03 13:36:45,401 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r3_c0.jpg
+2025-03-03 13:36:45,632 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r4_c0.jpg
+2025-03-03 13:36:45,853 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r4_c1.jpg
+2025-03-03 13:36:46,263 [INFO] __main__ - Processing table image: /topic-extraction/img_28.jpg, columns=two
+2025-03-03 13:36:49,648 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_28.jpg_r0_c0.jpg
+2025-03-03 13:36:49,911 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_28.jpg_r1_c0.jpg
+2025-03-03 13:36:50,133 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_28.jpg_r2_c0.jpg
+2025-03-03 13:36:50,350 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_28.jpg_r3_c0.jpg
+2025-03-03 13:36:50,760 [INFO] __main__ - Final subtopics JSON saved locally at /home/user/app/we/final_subtopics.json
+2025-03-03 13:36:51,086 [INFO] __main__ - GPU memory cleaned up.
+2025-03-03 13:36:51,093 [INFO] __main__ - Processing completed successfully.
+2025-03-03 14:05:17,866 [INFO] __main__ - Processing PDF: /home/user/app/input_output/a-level-pearson-mathematics-specification.pdf
+2025-03-03 14:05:18,700 [INFO] __main__ - Gemini returned subtopics: {'Paper 1 and Paper 2: Pure Mathematics': [11, 29], 'Paper 3: Statistics and Mechanics': [30, 40]}
+2025-03-03 14:05:18,702 [INFO] __main__ - Loaded 1135473 bytes from local file '/home/user/app/input_output/a-level-pearson-mathematics-specification.pdf'
+2025-03-03 14:05:19,046 [INFO] __main__ - Computed global offset: 4
+2025-03-03 14:05:19,047 [INFO] __main__ - Processing pages (0-based): [14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43]
+2025-03-03 14:05:52,370 [INFO] __main__ - GPU memory cleaned up.
+2025-03-03 14:10:28,391 [INFO] __main__ - Processing PDF: /home/user/app/input_output/a-level-pearson-mathematics-specification.pdf
+2025-03-03 14:10:29,161 [INFO] __main__ - Gemini returned subtopics: {'Paper 1 and Paper 2: Pure Mathematics': [11, 29], 'Paper 3: Statistics and Mechanics': [30, 40]}
+2025-03-03 14:10:29,162 [INFO] __main__ - Loaded 1135473 bytes from local file '/home/user/app/input_output/a-level-pearson-mathematics-specification.pdf'
+2025-03-03 14:10:29,484 [INFO] __main__ - Computed global offset: 4
+2025-03-03 14:10:29,484 [INFO] __main__ - Processing pages (0-based): [14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43]
+2025-03-03 14:11:29,432 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_1.jpg
+2025-03-03 14:11:31,185 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_2.jpg
+2025-03-03 14:11:31,702 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_3.jpg
+2025-03-03 14:11:32,212 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_4.jpg
+2025-03-03 14:11:32,763 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_5.jpg
+2025-03-03 14:11:33,406 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_6.jpg
+2025-03-03 14:11:33,898 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_7.jpg
+2025-03-03 14:11:34,395 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_8.jpg
+2025-03-03 14:11:34,878 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_9.jpg
+2025-03-03 14:11:35,417 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_10.jpg
+2025-03-03 14:11:35,838 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_11.jpg
+2025-03-03 14:11:36,385 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_12.jpg
+2025-03-03 14:11:36,971 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_13.jpg
+2025-03-03 14:11:37,669 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_14.jpg
+2025-03-03 14:11:38,314 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_15.jpg
+2025-03-03 14:11:38,926 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_16.jpg
+2025-03-03 14:11:39,484 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_17.jpg
+2025-03-03 14:11:39,846 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_18.jpg
+2025-03-03 14:11:40,381 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_19.jpg
+2025-03-03 14:11:40,979 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_20.jpg
+2025-03-03 14:11:41,538 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_21.jpg
+2025-03-03 14:11:42,104 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_22.jpg
+2025-03-03 14:11:42,640 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_23.jpg
+2025-03-03 14:11:43,153 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_24.jpg
+2025-03-03 14:11:43,663 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_25.jpg
+2025-03-03 14:11:44,172 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_26.jpg
+2025-03-03 14:11:44,677 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_27.jpg
+2025-03-03 14:11:45,255 [INFO] __main__ - Uploaded to S3: /topic-extraction/img_28.jpg
+2025-03-03 14:11:45,680 [INFO] __main__ - Classifying images to detect tables.
+2025-03-03 14:11:49,521 [INFO] __main__ - Processing table image: /topic-extraction/img_1.jpg, columns=three
+2025-03-03 14:11:52,761 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_1.jpg_r0_c0.jpg
+2025-03-03 14:11:52,762 [ERROR] __main__ - Error processing table image /topic-extraction/img_1.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:11:52,762 [INFO] __main__ - Processing table image: /topic-extraction/img_2.jpg, columns=three
+2025-03-03 14:11:56,000 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_2.jpg_r0_c0.jpg
+2025-03-03 14:11:56,000 [ERROR] __main__ - Error processing table image /topic-extraction/img_2.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:11:56,000 [INFO] __main__ - Processing table image: /topic-extraction/img_3.jpg, columns=three
+2025-03-03 14:11:59,555 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_3.jpg_r0_c0.jpg
+2025-03-03 14:11:59,555 [ERROR] __main__ - Error processing table image /topic-extraction/img_3.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:11:59,556 [INFO] __main__ - Processing table image: /topic-extraction/img_4.jpg, columns=three
+2025-03-03 14:12:02,696 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_4.jpg_r0_c0.jpg
+2025-03-03 14:12:02,697 [ERROR] __main__ - Error processing table image /topic-extraction/img_4.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:02,697 [INFO] __main__ - Processing table image: /topic-extraction/img_5.jpg, columns=three
+2025-03-03 14:12:06,308 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_5.jpg_r0_c0.jpg
+2025-03-03 14:12:06,309 [ERROR] __main__ - Error processing table image /topic-extraction/img_5.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:06,309 [INFO] __main__ - Processing table image: /topic-extraction/img_6.jpg, columns=three
+2025-03-03 14:12:10,140 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_6.jpg_r0_c0.jpg
+2025-03-03 14:12:10,141 [ERROR] __main__ - Error processing table image /topic-extraction/img_6.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:10,141 [INFO] __main__ - Processing table image: /topic-extraction/img_7.jpg, columns=three
+2025-03-03 14:12:13,713 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_7.jpg_r0_c0.jpg
+2025-03-03 14:12:13,713 [ERROR] __main__ - Error processing table image /topic-extraction/img_7.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:13,713 [INFO] __main__ - Processing table image: /topic-extraction/img_8.jpg, columns=three
+2025-03-03 14:12:17,306 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_8.jpg_r0_c0.jpg
+2025-03-03 14:12:17,306 [ERROR] __main__ - Error processing table image /topic-extraction/img_8.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:17,306 [INFO] __main__ - Processing table image: /topic-extraction/img_9.jpg, columns=three
+2025-03-03 14:12:21,354 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_9.jpg_r0_c0.jpg
+2025-03-03 14:12:21,354 [ERROR] __main__ - Error processing table image /topic-extraction/img_9.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:21,355 [INFO] __main__ - Processing table image: /topic-extraction/img_10.jpg, columns=three
+2025-03-03 14:12:24,668 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_10.jpg_r0_c0.jpg
+2025-03-03 14:12:24,668 [ERROR] __main__ - Error processing table image /topic-extraction/img_10.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:24,668 [INFO] __main__ - Processing table image: /topic-extraction/img_11.jpg, columns=two
+2025-03-03 14:12:27,914 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_11.jpg_r0_c0.jpg
+2025-03-03 14:12:27,914 [ERROR] __main__ - Error processing table image /topic-extraction/img_11.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:27,915 [INFO] __main__ - Processing table image: /topic-extraction/img_12.jpg, columns=three
+2025-03-03 14:12:31,345 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_12.jpg_r0_c0.jpg
+2025-03-03 14:12:31,346 [ERROR] __main__ - Error processing table image /topic-extraction/img_12.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:31,346 [INFO] __main__ - Processing table image: /topic-extraction/img_13.jpg, columns=three
+2025-03-03 14:12:34,536 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_13.jpg_r0_c0.jpg
+2025-03-03 14:12:34,536 [ERROR] __main__ - Error processing table image /topic-extraction/img_13.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:34,536 [INFO] __main__ - Processing table image: /topic-extraction/img_14.jpg, columns=three
+2025-03-03 14:12:39,055 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_14.jpg_r0_c0.jpg
+2025-03-03 14:12:39,056 [ERROR] __main__ - Error processing table image /topic-extraction/img_14.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:39,056 [INFO] __main__ - Processing table image: /topic-extraction/img_15.jpg, columns=three
+2025-03-03 14:12:43,762 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_15.jpg_r0_c0.jpg
+2025-03-03 14:12:43,763 [ERROR] __main__ - Error processing table image /topic-extraction/img_15.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:43,763 [INFO] __main__ - Processing table image: /topic-extraction/img_16.jpg, columns=three
+2025-03-03 14:12:48,110 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_16.jpg_r0_c0.jpg
+2025-03-03 14:12:48,110 [ERROR] __main__ - Error processing table image /topic-extraction/img_16.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:48,110 [INFO] __main__ - Processing table image: /topic-extraction/img_17.jpg, columns=three
+2025-03-03 14:12:52,283 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_17.jpg_r0_c0.jpg
+2025-03-03 14:12:52,283 [ERROR] __main__ - Error processing table image /topic-extraction/img_17.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:52,284 [INFO] __main__ - Processing table image: /topic-extraction/img_18.jpg, columns=three
+2025-03-03 14:12:53,748 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_18.jpg_r0_c0.jpg
+2025-03-03 14:12:53,748 [ERROR] __main__ - Error processing table image /topic-extraction/img_18.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:53,749 [INFO] __main__ - Processing table image: /topic-extraction/img_19.jpg, columns=three
+2025-03-03 14:12:57,191 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_19.jpg_r0_c0.jpg
+2025-03-03 14:12:57,191 [ERROR] __main__ - Error processing table image /topic-extraction/img_19.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:12:57,192 [INFO] __main__ - Processing table image: /topic-extraction/img_20.jpg, columns=three
+2025-03-03 14:13:01,859 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_20.jpg_r0_c0.jpg
+2025-03-03 14:13:01,862 [ERROR] __main__ - Error processing table image /topic-extraction/img_20.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:01,862 [INFO] __main__ - Processing table image: /topic-extraction/img_21.jpg, columns=three
+2025-03-03 14:13:05,555 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_21.jpg_r0_c0.jpg
+2025-03-03 14:13:05,556 [ERROR] __main__ - Error processing table image /topic-extraction/img_21.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:05,556 [INFO] __main__ - Processing table image: /topic-extraction/img_22.jpg, columns=three
+2025-03-03 14:13:10,335 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_22.jpg_r0_c0.jpg
+2025-03-03 14:13:10,336 [ERROR] __main__ - Error processing table image /topic-extraction/img_22.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:10,336 [INFO] __main__ - Processing table image: /topic-extraction/img_23.jpg, columns=three
+2025-03-03 14:13:13,658 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_23.jpg_r0_c0.jpg
+2025-03-03 14:13:13,658 [ERROR] __main__ - Error processing table image /topic-extraction/img_23.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:13,659 [INFO] __main__ - Processing table image: /topic-extraction/img_24.jpg, columns=three
+2025-03-03 14:13:17,286 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_24.jpg_r0_c0.jpg
+2025-03-03 14:13:17,287 [ERROR] __main__ - Error processing table image /topic-extraction/img_24.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:17,287 [INFO] __main__ - Processing table image: /topic-extraction/img_25.jpg, columns=two
+2025-03-03 14:13:20,933 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_25.jpg_r0_c0.jpg
+2025-03-03 14:13:20,933 [ERROR] __main__ - Error processing table image /topic-extraction/img_25.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:20,933 [INFO] __main__ - Processing table image: /topic-extraction/img_26.jpg, columns=three
+2025-03-03 14:13:23,943 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_26.jpg_r0_c0.jpg
+2025-03-03 14:13:23,943 [ERROR] __main__ - Error processing table image /topic-extraction/img_26.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:23,944 [INFO] __main__ - Processing table image: /topic-extraction/img_27.jpg, columns=three
+2025-03-03 14:13:27,633 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_27.jpg_r0_c0.jpg
+2025-03-03 14:13:27,633 [ERROR] __main__ - Error processing table image /topic-extraction/img_27.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:27,633 [INFO] __main__ - Processing table image: /topic-extraction/img_28.jpg, columns=two
+2025-03-03 14:13:31,099 [INFO] __main__ - Uploaded to S3: /topic-extraction/cells/table_s3_img_28.jpg_r0_c0.jpg
+2025-03-03 14:13:31,100 [ERROR] __main__ - Error processing table image /topic-extraction/img_28.jpg: 'TableExtractor' object has no attribute 'crop_cell'
+2025-03-03 14:13:31,102 [INFO] __main__ - Final subtopics JSON saved locally at /home/user/app/we/final_subtopics.json
+2025-03-03 14:13:31,434 [INFO] __main__ - GPU memory cleaned up.
+2025-03-03 14:13:31,442 [INFO] __main__ - Processing completed successfully.

we/final_subtopics.json

@@ -0,0 +1 @@
+[]