149 lines
4.8 KiB
Python
149 lines
4.8 KiB
Python
#!/usr/bin/env python3
|
|
"""Extract images from docx with context (surrounding paragraphs).
|
|
|
|
Parses document.xml to find <w:drawing> elements and extracts:
|
|
- Image file path (from word/media/)
|
|
- Relationship ID
|
|
- Context: text from paragraphs before and after the image
|
|
|
|
Output: JSON with image list and context info.
|
|
"""
|
|
|
|
import json
|
|
import sys
|
|
import zipfile
|
|
import xml.etree.ElementTree as ET
|
|
from pathlib import Path
|
|
|
|
|
|
# Namespaces
|
|
W_NS = "http://schemas.openxmlformats.org/wordprocessingml/2006/main"
|
|
R_NS = "http://schemas.openxmlformats.org/officeDocument/2006/relationships"
|
|
A_NS = "http://schemas.openxmlformats.org/drawingml/2006/main"
|
|
|
|
NS = {
|
|
"w": W_NS,
|
|
"wp": "http://schemas.openxmlformats.org/drawingml/2006/wordprocessingDrawing",
|
|
"a": A_NS,
|
|
"pic": "http://schemas.openxmlformats.org/drawingml/2006/picture",
|
|
"r": R_NS,
|
|
}
|
|
|
|
|
|
def extract_docx_images(docx_path: str, output_dir: str) -> dict:
|
|
"""Extract images from docx with context information.
|
|
|
|
Args:
|
|
docx_path: Path to the .docx file
|
|
output_dir: Directory to extract images to
|
|
|
|
Returns:
|
|
Dict with 'images' list containing path, relId, contextBefore, contextAfter
|
|
"""
|
|
docx = Path(docx_path)
|
|
out = Path(output_dir)
|
|
out.mkdir(parents=True, exist_ok=True)
|
|
|
|
if not docx.exists():
|
|
return {"error": f"File not found: {docx_path}", "images": []}
|
|
|
|
try:
|
|
with zipfile.ZipFile(docx, "r") as zf:
|
|
# Extract all media files
|
|
media_files = [f for f in zf.namelist() if f.startswith("word/media/")]
|
|
for mf in media_files:
|
|
zf.extract(mf, out)
|
|
|
|
# Parse document.xml for image positions
|
|
with zf.open("word/document.xml") as f:
|
|
doc_xml = f.read()
|
|
|
|
# Parse relationships
|
|
with zf.open("word/_rels/document.xml.rels") as f:
|
|
rels_xml = f.read()
|
|
except Exception as e:
|
|
return {"error": f"Failed to read docx: {e}", "images": []}
|
|
|
|
# Build relId -> media file mapping
|
|
rels_root = ET.fromstring(rels_xml)
|
|
rel_map = {}
|
|
for rel in rels_root:
|
|
rel_id = rel.get("Id")
|
|
target = rel.get("Target") # e.g., "media/image1.png"
|
|
if rel_id and target and target.startswith("media/"):
|
|
rel_map[rel_id] = target
|
|
|
|
# Parse document to find images with context
|
|
doc_root = ET.fromstring(doc_xml)
|
|
body = doc_root.find(".//w:body", NS)
|
|
if body is None:
|
|
return {"error": "No body found in document", "images": []}
|
|
|
|
paragraphs = list(body.findall("w:p", NS))
|
|
|
|
images = []
|
|
for idx, para in enumerate(paragraphs):
|
|
# Find all drawings in this paragraph
|
|
drawings = para.findall(".//w:drawing", NS)
|
|
if not drawings:
|
|
continue
|
|
|
|
# Get context: text from previous and next non-empty paragraphs
|
|
context_before = ""
|
|
for i in range(idx - 1, -1, -1):
|
|
text = "".join(t.text or "" for t in paragraphs[i].findall(".//w:t", NS))
|
|
if text.strip():
|
|
context_before = text.strip()
|
|
break
|
|
|
|
context_after = ""
|
|
for i in range(idx + 1, len(paragraphs)):
|
|
text = "".join(t.text or "" for t in paragraphs[i].findall(".//w:t", NS))
|
|
if text.strip():
|
|
context_after = text.strip()
|
|
break
|
|
|
|
# Extract image info from each drawing
|
|
for drawing in drawings:
|
|
# Find the blip reference
|
|
blip = drawing.find(".//a:blip", NS)
|
|
if blip is None:
|
|
continue
|
|
|
|
rel_id = blip.get(f"{{{R_NS}}}embed")
|
|
if not rel_id:
|
|
continue
|
|
|
|
media_target = rel_map.get(rel_id)
|
|
if not media_target:
|
|
continue
|
|
|
|
# Copy image to output directory
|
|
# zip extracts to word/media/, rels target is media/
|
|
src_path = out / "word" / media_target
|
|
if src_path.exists():
|
|
filename = Path(media_target).name
|
|
dest_path = out / filename
|
|
if src_path != dest_path:
|
|
import shutil
|
|
shutil.copy2(str(src_path), str(dest_path))
|
|
|
|
images.append({
|
|
"filename": filename,
|
|
"path": str(dest_path),
|
|
"relId": rel_id,
|
|
"contextBefore": context_before,
|
|
"contextAfter": context_after,
|
|
})
|
|
|
|
return {"images": images}
|
|
|
|
|
|
if __name__ == "__main__":
|
|
if len(sys.argv) != 3:
|
|
print("Usage: extract-docx-images.py <docx_path> <output_dir>")
|
|
sys.exit(1)
|
|
|
|
result = extract_docx_images(sys.argv[1], sys.argv[2])
|
|
print(json.dumps(result, ensure_ascii=False, indent=2))
|