Jarian Cottingham
c0fa9148b7
- mcp/server.py: stdlib-only MCP stdio server exposing ocr_image tool - ocr-posted.py: CLI equivalent (DB lookup + OCR via llama-server) - run-ocr-server.sh: llama-server launcher with tuned sampling flags - skill/ocr-image: opencode skill teaching agents when/how to call it - README with full install instructions
77 lines
2.6 KiB
Python
Executable File
77 lines
2.6 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
"""Pull an image posted into opencode from opencode.db and OCR it via local Unlimited-OCR.
|
|
|
|
Usage:
|
|
ocr-posted.py [latest | <part_id> | session:<session_id>] [prompt] [save_path]
|
|
|
|
Defaults: latest image part in DB, prompt "Free OCR."
|
|
Env: OPENCODE_DB, OCR_URL
|
|
"""
|
|
import base64, json, os, sqlite3, sys, urllib.request
|
|
|
|
DB = os.environ.get(
|
|
'OPENCODE_DB',
|
|
os.path.expanduser('~/.local/share/opencode/opencode.db'))
|
|
OCR_URL = os.environ.get('OCR_URL', 'http://127.0.0.1:11436/v1/chat/completions')
|
|
|
|
|
|
def pick(db, arg):
|
|
cur = db.cursor()
|
|
if arg and arg != 'latest':
|
|
if arg.startswith('session:'):
|
|
cur.execute(
|
|
"SELECT id, data FROM part WHERE session_id=? "
|
|
"AND data LIKE '%\"type\":\"file\"%' AND data LIKE '%data:image%' "
|
|
"ORDER BY time_created DESC LIMIT 1",
|
|
(arg[8:],))
|
|
else:
|
|
cur.execute("SELECT id, data FROM part WHERE id=?", (arg,))
|
|
else:
|
|
cur.execute(
|
|
"SELECT id, data FROM part WHERE data LIKE '%\"type\":\"file\"%' "
|
|
"AND data LIKE '%data:image%' ORDER BY time_created DESC LIMIT 1")
|
|
return cur.fetchone()
|
|
|
|
|
|
def ocr(data_url, prompt, max_tokens=4096):
|
|
body = json.dumps({
|
|
"model": "Unlimited-OCR",
|
|
"temperature": 0,
|
|
"repeat_penalty": 1.2,
|
|
"max_tokens": max_tokens,
|
|
"messages": [{"role": "user", "content": [
|
|
{"type": "image_url", "image_url": {"url": data_url}},
|
|
{"type": "text", "text": prompt},
|
|
]}],
|
|
}).encode()
|
|
req = urllib.request.Request(
|
|
OCR_URL, data=body, headers={"Content-Type": "application/json"})
|
|
with urllib.request.urlopen(req, timeout=1800) as r:
|
|
return json.load(r)['choices'][0]['message']['content']
|
|
|
|
|
|
def main():
|
|
arg = sys.argv[1] if len(sys.argv) > 1 else 'latest'
|
|
prompt = sys.argv[2] if len(sys.argv) > 2 else 'Free OCR.'
|
|
save = sys.argv[3] if len(sys.argv) > 3 else None
|
|
db = sqlite3.connect(DB, timeout=10)
|
|
row = pick(db, arg)
|
|
if not row:
|
|
print('no image part found', file=sys.stderr)
|
|
sys.exit(1)
|
|
pid, data = row
|
|
d = json.loads(data)
|
|
url = d.get('url', '')
|
|
if not url.startswith('data:image'):
|
|
print('part %s has no inline image data' % pid, file=sys.stderr)
|
|
sys.exit(1)
|
|
if save:
|
|
open(save, 'wb').write(base64.b64decode(url.split(',', 1)[1]))
|
|
print('saved image to %s' % save, file=sys.stderr)
|
|
print('part: %s file: %s' % (pid, d.get('filename')), file=sys.stderr)
|
|
print(ocr(url, prompt))
|
|
|
|
|
|
if __name__ == '__main__':
|
|
main()
|