pdf-extraction-fallbacks
Multi-fallback PDF/text extraction with early failure detection and sequential tool fallbacks
- License 许可
- MIT
- 所属仓库
- HKUDS/AgentPilot
- 最近更新
- 2026-11-12
- Stars
- ★ 8.0k
- Forks
- 944
01 前言 Front Matter
- name
- pdf-extraction-fallbacks
- description
- Multi-fallback PDF/text extraction with early failure detection and sequential tool fallbacks
02 资源文件 1
- 其他.skill_id38 B
03 SKILL.md 正文
PDF Extraction with Multi-Fallback Strategy
Purpose
When extracting text from PDFs (especially regulatory documents, handbooks, or protected content), single-method approaches often fail due to JavaScript protection, CORS restrictions, encoding issues, or corrupted downloads. This skill provides a robust multi-fallback workflow that detects failures early and tries sequential extraction methods.
Core Pattern
- Download with validation - Check file size and content sanity immediately
- Sequential extraction attempts - Try multiple tools in order of reliability
- Early failure detection - Don't proceed with obviously corrupt files
- Document fallback path - Log which method succeeded for future reference
Step-by-Step Instructions
Step 1: Download and Validate
Before attempting extraction, validate the downloaded file:
# Download the PDF
curl -L -o document.pdf "$URL"
# Check file size (reject if < 1KB - likely error page)
FILE_SIZE=$(stat -f%z document.pdf 2>/dev/null || stat -c%s document.pdf 2>/dev/null)
if [ "$FILE_SIZE" -lt 1024 ]; then
echo "ERROR: File too small ($FILE_SIZE bytes) - likely not a valid PDF"
# Check if it's an HTML error page
head -c 200 document.pdf | grep -i "<html\|<!doctype\|error\|access denied" && \
echo "Detected HTML error page instead of PDF"
exit 1
fi
# Check PDF magic bytes
HEAD_BYTES=$(head -c 4 document.pdf)
if [ "$HEAD_BYTES" != "%PDF" ]; then
echo "ERROR: File does not start with PDF magic bytes"
head -c 100 document.pdf
exit 1
fi
Step 2: Primary Extraction (pdftotext)
# Try pdftotext first (fastest, most reliable for simple PDFs)
if command -v pdftotext &> /dev/null; then
pdftotext -layout document.pdf output.txt 2>/dev/null
if [ -s output.txt ]; then
WORD_COUNT=$(wc -w < output.txt)
if [ "$WORD_COUNT" -gt 50 ]; then
echo "SUCCESS: pdftotext extracted $WORD_COUNT words"
exit 0
fi
fi
fi
Step 3: Fallback 1 (PyMuPDF/fitz)
# Try PyMuPDF - handles more complex PDFs
import fitz # pymupdf
try:
doc = fitz.open("document.pdf")
text = ""
for page in doc:
text += page.get_text()
if len(text.strip()) > 500: # Sanity check
with open("output.txt", "w") as f:
f.write(text)
print(f"SUCCESS: PyMuPDF extracted {len(text)} characters")
else:
print("WARNING: PyMuPDF extraction too short, trying next method")
except Exception as e:
print(f"PyMuPDF failed: {e}")
Step 4: Fallback 2 (pdfplumber)
# Try pdfplumber - better for tables and structured content
import pdfplumber
try:
text = ""
with pdfplumber.open("document.pdf") as pdf:
for page in pdf.pages:
page_text = page.extract_text()
if page_text:
text += page_text + "\n"
if len(text.strip()) > 500:
with open("output.txt", "w") as f:
f.write(text)
print(f"SUCCESS: pdfplumber extracted {len(text)} characters")
else:
print("WARNING: pdfplumber extraction too short")
except Exception as e:
print(f"pdfplumber failed: {e}")
Step 5: Handle JavaScript-Protected Pages
If all methods fail, the PDF may be JavaScript-protected:
# Check for JavaScript in PDF
import fitz
doc = fitz.open("document.pdf")
has_js = False
for page in doc:
if page.get_java_script():
has_js = True
break
if has_js:
print("WARNING: PDF contains JavaScript - may be protected")
# Try rendering pages as images and OCR (requires additional tools)
# Or try alternative download source
Step 6: Alternative Sources
If the primary URL fails:
- Try alternative domains (e.g.,
.govmirrors, archive.org) - Check if the document is available via API
- Look for HTML version of the same content
- Search for the document title + "pdf" to find mirrors
Decision Tree
Download PDF
│
├─→ File < 1KB? → REJECT (likely error page)
├─→ No %PDF header? → REJECT (not a PDF)
│
└─→ Valid PDF
│
├─→ pdftotext → >50 words? → SUCCESS
│ └─→ Try next
│
├─→ PyMuPDF → >500 chars? → SUCCESS
│ └─→ Try next
│
├─→ pdfplumber → >500 chars? → SUCCESS
│ └─→ Try next
│
└─→ All failed → Check for JS protection, try alternative sources
Common Failure Modes
| Symptom | Likely Cause | Solution |
|---|---|---|
| File < 100 bytes | JavaScript error page | Check CORS, try different user-agent |
| File ~1-5KB | HTML error/warning page | Parse HTML for actual PDF link |
| pdftotext returns empty | Encrypted/protected PDF | Try PyMuPDF with password handling |
| Garbled text output | Encoding issue | Try pdfplumber, specify encoding |
| Extraction very short | Images-only PDF | Need OCR (tesseract) |
Example Complete Workflow Script
Save as extract_pdf_robust.sh:
#!/bin/bash
set -e
URL="$1"
OUTPUT="${2:-output.txt}"
TEMP_PDF="temp_download.pdf"
echo "Downloading from: $URL"
curl -L -A "Mozilla/5.0" -o "$TEMP_PDF" "$URL"
# Validate
SIZE=$(stat -c%s "$TEMP_PDF" 2>/dev/null || stat -f%z "$TEMP_PDF")
echo "Downloaded: $SIZE bytes"
if [ "$SIZE" -lt 1024 ]; then
echo "ERROR: File too small - checking content..."
head -200 "$TEMP_PDF"
exit 1
fi
if ! head -c 4 "$TEMP_PDF" | grep -q "%PDF"; then
echo "ERROR: Not a valid PDF file"
head -100 "$TEMP_PDF"
exit 1
fi
# Try extraction methods
python3 << 'PYTHON'
import sys
import fitz
import pdfplumber
pdf_path = "temp_download.pdf"
output_path = "output.txt"
# Method 1: PyMuPDF
try:
doc = fitz.open(pdf_path)
text = "".join(page.get_text() for page in doc)
if len(text.strip()) > 500:
with open(output_path, "w") as f:
f.write(text)
print(f"PyMuPDF: {len(text)} chars")
sys.exit(0)
except Exception as e:
print(f"PyMuPDF failed: {e}")
# Method 2: pdfplumber
try:
text = ""
with pdfplumber.open(pdf_path) as pdf:
for page in pdf.pages:
txt = page.extract_text()
if txt:
text += txt + "\n"
if len(text.strip()) > 500:
with open(output_path, "w") as f:
f.write(text)
print(f"pdfplumber: {len(text)} chars")
sys.exit(0)
except Exception as e:
print(f"pdfplumber failed: {e}")
print("All extraction methods failed")
sys.exit(1)
PYTHON
Best Practices
- Always validate - Never assume a download succeeded
- Log the path taken - Record which method worked for debugging
- Set reasonable thresholds - 50 words / 500 chars minimum for "success"
- Keep raw PDF - Don't delete the original until extraction is confirmed
- Retry with variations - Different user-agents, referer headers, or mirrors
Dependencies
curl- For downloadingpoppler-utils(pdftotext) - Optional, fast extractionPyMuPDF(fitz) -pip install pymupdfpdfplumber-pip install pdfplumber
04 同仓库更多 Skills
agentpilot
Set up AgentPilot and use its single entry to prepare missing inputs and run configured RSI experiments. Use when the user requests AgentPilot installation or work with AgentPilot.
remember
Review auto-memory entries and propose promotions to OPENSPACE.md, OPENSPACE.local.md, or retained auto-memory topics; detect duplicates, stale facts, conflicts, and ambiguous entries before making changes.
skill-discovery
Search reusable OpenSpace skills locally, and browse cloud packages step by step when a cloud skill may help. Reusing proven skills saves tokens, improves reliability, and extends your capabilities beyond built-in tools.
delegate-task
Delegate tasks to OpenSpace — a full-stack autonomous worker for coding, DevOps, web research, and desktop automation, backed by an extensive MCP tool and skill library. Skills auto-improve through use, reducing token consumption over time. A cloud community lets agents share and collectively evolve reusable skills.
write-special-chars-ts-check
Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, plus a reliable TypeScript type-checking step that filters output to only the newly-written files to avoid being misled by pre-existing errors.
write-special-chars-readback
Workaround for write_file failures caused by special characters (apostrophes, backticks, template literals) using a Python heredoc script, with mandatory post-write verification via read_file and cleanup of /tmp helper scripts.