feat: add vqa pipeline

[ChenZiHong-Gavin]

This PRintroduces significant improvements to the handling of multi-modal (text, image, table, equation) documents and their integration into the knowledge graph generation pipeline. The changes include a refactor of the document insertion workflow to separately process text and multi-modal documents, the addition of filtering and validation for input data, and updates to key data structures and configuration files to better support multi-modal processing. Logging has also been improved for better debugging and traceability.

[Copilot]

Pull Request Overview

This PR adds a VQA (Visual Question Answering) pipeline by introducing support for multimodal document processing. The key changes include:

• Adding a centralized filter method in BaseReader to validate and filter text, image, table, and equation entries
• Updating all reader implementations to use this common filtering logic
• Providing a new VQA demo JSON file with scientific paper content including text, images, and tables

Reviewed Changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
resources/input_examples/vqa_demo.json	New demo file containing scientific paper data with text, images, tables, and equations
graphgen/bases/base_reader.py	Added filter method to validate content and check image existence
graphgen/models/reader/txt_reader.py	Updated to call filter method on results
graphgen/models/reader/pdf_reader.py	Updated to call filter method and removed duplicate filtering logic
graphgen/models/reader/jsonl_reader.py	Updated validation logic and added filter call
graphgen/models/reader/json_reader.py	Updated validation logic and added filter call
graphgen/models/reader/csv_reader.py	Updated validation logic and added filter call
graphgen/configs/vqa_config.yaml	Changed input file from PDF demo to VQA JSON demo
graphgen/graphgen.py	Added debug print statement

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Pull Request Overview

Copilot reviewed 21 out of 27 changed files in this pull request and generated 9 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The _image_exists function makes a network request for every URL with a 3-second timeout. For documents with many images, this could significantly slow down the filtering process. Consider implementing caching or batch validation to improve performance.

[Copilot]

Pull Request Overview

Copilot reviewed 37 out of 43 changed files in this pull request and generated 4 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Direct modification of dictionary during iteration is error-prone. The deletion of keys from item should happen before appending to results, or a filtered copy should be created instead. Consider using dictionary comprehension: item = {k: v for k, v in item.items() if k not in ('page_idx', 'bbox', 'text_level')}.

[Copilot]

Using df.iterrows() is inefficient for pandas DataFrames. This validation should be performed using vectorized operations or at least df.itertuples() for better performance. Consider: text_rows = df[df['type'] == 'text']; if text_rows[self.text_column].isna().any(): raise ValueError(...).

Suggested change

	for _, row in df.iterrows():
	if row.get("type") == "text" and self.text_column not in row:
	raise ValueError(
	f"Missing '{self.text_column}' in document: {row.to_dict()}"
	)
	text_rows = df[df["type"] == "text"]
	if self.text_column not in df.columns or text_rows[self.text_column].isna().any():
	# Find the first offending row for error message
	missing_row = text_rows[text_rows[self.text_column].isna()].iloc[0] if self.text_column in df.columns and not text_rows.empty else None
	raise ValueError(
	f"Missing '{self.text_column}' in document: {missing_row.to_dict() if missing_row is not None else 'Column not found'}"
	)

[Copilot]

The _image_exists function makes a blocking HTTP request for each image URL, which will severely impact performance when processing documents with many images. Consider adding async support or implementing batch validation with connection pooling.

[Copilot]

Pull Request Overview

Copilot reviewed 40 out of 46 changed files in this pull request and generated 3 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

Pull Request Overview

Copilot reviewed 40 out of 46 changed files in this pull request and generated 6 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}