メニュー
Extract text, tables, and structured information from PDF documents using pdfplumber, PyPDF2, or pdftotext command-line tools.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
| name | pdf-reading |
| description | Extract text, tables, and structured information from PDF documents using pdfplumber, PyPDF2, or pdftotext command-line tools. |
This skill helps agents extract information from PDF documents.
The environment has these PDF tools pre-installed:
pdfplumber - Python library for precise text/table extractionPyPDF2 - Python library for PDF manipulationpdftotext - Command-line tool from poppler-utils# Extract all text from a PDF
pdftotext /root/artifacts/paper.pdf -
# Extract specific pages
pdftotext -f 1 -l 3 /root/artifacts/paper.pdf -
# Extract to a file
pdftotext /root/artifacts/paper.pdf /tmp/paper.txt
import pdfplumber
from pathlib import Path
def extract_pdf_text(pdf_path: str) -> str:
"""Extract all text from a PDF file."""
text_parts = []
with pdfplumber.open(pdf_path) as pdf:
for page in pdf.pages:
text = page.extract_text()
if text:
text_parts.append(text)
return "\n\n".join(text_parts)
# Usage
text = extract_pdf_text("/root/artifacts/paper.pdf")
print(text)
import re
def find_commands(text: str) -> list:
"""Extract shell commands from text."""
# Look for common command patterns
patterns = [
r'docker run[^\n]+',
r'\$[^\n]+',
r'--package=[^\s]+\s+--version=[^\s]+',
]
commands = []
for pattern in patterns:
commands.extend(re.findall(pattern, text))
return commands
text = extract_pdf_text("/root/artifacts/paper.pdf")
commands = find_commands(text)
import pdfplumber
def extract_tables(pdf_path: str) -> list:
"""Extract all tables from a PDF."""
tables = []
with pdfplumber.open(pdf_path) as pdf:
for i, page in enumerate(pdf.pages):
page_tables = page.extract_tables()
for table in page_tables:
tables.append({
"page": i + 1,
"data": table
})
return tables
import re
def find_package_info(text: str) -> list:
"""Find npm package references (name@version)."""
# Match patterns like node-rsync@1.0.3
pattern = r'([a-z0-9-]+)@(\d+\.\d+\.\d+)'
matches = re.findall(pattern, text.lower())
return [{"name": m[0], "version": m[1]} for m in matches]
Check available PDFs first:
ls -la /root/artifacts/
Preview before full extraction:
pdftotext /root/artifacts/paper.pdf - | head -100
Handle multi-column layouts: pdfplumber handles them better than pdftotext
For structured data: Look for JSON blocks in the text:
import json
import re
json_blocks = re.findall(r'\{[^{}]*\}', text)
for block in json_blocks:
try:
data = json.loads(block)
print(data)
except json.JSONDecodeError:
pass
Interact with research artifacts running in separate Docker containers via artifact-runner. Execute commands through HTTP API, read files, and verify artifact functionality.
Evaluate research artifacts running in separate Docker containers via artifact-runner. Access artifacts through HTTP API, execute commands, read files, and analyze PDFs.
Interact with artifact containers via HTTP API for paper evaluation tasks. Execute commands, read files, and list directories in remote artifact environments.
Evaluate research artifacts against NDSS badge criteria (Available, Functional, Reproduced) by checking DOI, documentation, exercisability, and reproducibility requirements.
Calculate protein dihedral angle (phi/psi) from atomic coordinates using the torsion formula. Use when computing backbone dihedral angles from atomic positions or validating protein geometry.
Analyze DNA/RNA/protein sequences: type determination, GC content calculation, reading frame translation, and longest ORF finding. Use when users ask about sequence translation, GC content, open reading frames, or residue counting.