| import json |
| from typing import Dict |
|
|
| import pandas as pd |
|
|
| from src.document_schema import DocumentIntelligenceReport |
|
|
|
|
| def report_to_dict(report: DocumentIntelligenceReport) -> Dict: |
| """Convert a Pydantic report into a regular dictionary.""" |
| return report.model_dump() |
|
|
|
|
| def report_to_json(report: DocumentIntelligenceReport) -> str: |
| """Convert the full report into formatted JSON.""" |
| return json.dumps(report_to_dict(report), indent=2) |
|
|
|
|
| def chunks_to_dataframe(report: DocumentIntelligenceReport) -> pd.DataFrame: |
| """Convert document chunks into a dataframe.""" |
| return pd.DataFrame([chunk.model_dump() for chunk in report.chunks]) |
|
|
|
|
| def findings_to_dataframe(report: DocumentIntelligenceReport) -> pd.DataFrame: |
| """Convert source-linked structured findings into a dataframe.""" |
| rows = [] |
| for finding in report.summary.findings: |
| rows.append( |
| { |
| "finding_id": finding.finding_id, |
| "category": finding.category.replace("_", " ").title(), |
| "finding": finding.text, |
| "page": finding.page_number, |
| "chunk": finding.chunk_id, |
| } |
| ) |
| return pd.DataFrame(rows) |
|
|
|
|
| def metadata_to_dataframe(report: DocumentIntelligenceReport) -> pd.DataFrame: |
| """Convert metadata into a dataframe.""" |
| return pd.DataFrame([report.metadata.model_dump()]) |
|
|
|
|
| def summary_to_dataframe(report: DocumentIntelligenceReport) -> pd.DataFrame: |
| """Convert summary fields into a simple dataframe.""" |
| summary = report.summary.model_dump(exclude={"findings"}) |
|
|
| rows = [] |
| for field, value in summary.items(): |
| if isinstance(value, list): |
| display_value = "\n".join(value) |
| else: |
| display_value = value |
|
|
| rows.append( |
| { |
| "field": field, |
| "value": display_value, |
| } |
| ) |
|
|
| return pd.DataFrame(rows) |
|
|
