PDF Reader

PDF Reader

Extract text from PDF files for text manipulation, search, and reference.

Quick Start

Extract all text from a PDF:

python scripts/extract_text.py document.pdf

Save to file:

python scripts/extract_text.py document.pdf -o .tmp/output.txt

Extract specific pages:

python scripts/extract_text.py document.pdf -p 1-10 -o .tmp/pages.txt

Workflow

1. Extract text → Run scripts/extract_text.py
2. Process output → Text is now searchable, editable, quotable
3. Reference content → Use extracted text for analysis or response

Script Options

extract_text.py <pdf_path> [options]

Options:
  -o, --output FILE      Save to file (default: print to stdout)
  -m, --method METHOD    auto|pdfplumber|pymupdf|pdfminer (default: auto)
  -p, --pages RANGE      Page range: "1-5" or "1,3,5" (default: all)
  --preserve-layout      Keep spatial arrangement of text
  --json                 Output with metadata (page sizes, method used)

Method Selection

Examples

Extract and search

python scripts/extract_text.py report.pdf | grep -i "revenue"

Extract tables (use pdfplumber)

python scripts/extract_text.py data.pdf -m pdfplumber --json -o .tmp/data.json

Specific pages with layout

python scripts/extract_text.py book.pdf -p 50-55 --preserve-layout -o .tmp/chapter.txt

Dependencies

At least one library required:

pip install pdfplumber pymupdf pdfminer.six

For detailed library comparison, see references/pdf_libraries.md.

Troubleshooting

Empty output?

• PDF may be scanned/image-based → try --method pymupdf (has OCR)
• Check if PDF is password-protected

Garbled text?

• Try different method: -m pdfminer
• PDF may have non-standard font encoding

Tables not formatted?

• Use -m pdfplumber --json for structured output
• Consider --preserve-layout flag

AGI Framework Integration

Qdrant Memory Integration

Before executing complex tasks with this skill:

python3 execution/memory_manager.py auto --query "<task summary>"

Decision Tree:

• Cache hit? Use cached response directly — no need to re-process.
• Memory match? Inject context_chunks into your reasoning.
• No match? Proceed normally, then store results:

python3 execution/memory_manager.py store \
  --content "Description of what was decided/solved" \
  --type decision \
  --tags pdf-reader <relevant-tags>

Note: Storing automatically updates both Vector (Qdrant) and Keyword (BM25) indices.

Agent Team Collaboration- Strategy: This skill communicates via the shared memory system.

• Orchestration: Invoked by orchestrator via intelligent routing.
• Context Sharing: Always read previous agent outputs from memory before starting.

Local LLM Support

When available, use local Ollama models for embedding and lightweight inference:

• Embeddings: nomic-embed-text via Qdrant memory system
• Lightweight analysis: Local models reduce API costs for repetitive patterns