orbitalo/homelab-brain
1
0
Fork 0
Code Issues 37 Pull requests Projects Releases Packages Wiki Activity Actions

fix: Ollama native API nutzen statt OpenAI-compat für Enricher

Browse source 
Qwen3 via /v1/chat/completions verbraucht max_tokens fürs
Reasoning und liefert leeren Content. Umstellung auf /api/chat
mit think=false löst das Problem.
This commit is contained in:
orbitalo 2026-03-27 12:36:38 +00:00
parent d8e3207259
commit caa2883a66
1 changed files with 9 additions and 6 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
homelab-ai-bot/savetv_enrich.py
View file

@ -58,25 +58,28 @@ def _is_enriched(entry: dict) -> bool:
def _call_ollama(prompt: str, model: str = MODEL) -> str: def _call_ollama(prompt: str, model: str = MODEL) -> str:
"""Ruft Ollama via native /api/chat auf (kein OpenAI-compat)."""
payload = { payload = {
"model": model, "model": model,
"messages": [ "messages": [
{"role": "system", "content": ( {"role": "system", "content": (
"Du bist eine Filmdatenbank. Antworte NUR mit validem JSON, " "Du bist eine Filmdatenbank. Antworte NUR mit validem JSON, "
"kein Markdown, keine Erklärungen. /no_think" "kein Markdown, keine Erklärungen."
)}, )},
{"role": "user", "content": prompt + " /no_think"}, {"role": "user", "content": prompt},
], ],
"max_tokens": 800,
"stream": False, "stream": False,
"think": False,
"options": {"num_predict": 1024},
} }
try: try:
r = requests.post( r = requests.post(
f"{OLLAMA_BASE}/v1/chat/completions", f"{OLLAMA_BASE}/api/chat",
json=payload, timeout=120, json=payload, timeout=180,
) )
r.raise_for_status() r.raise_for_status()
text = r.json()["choices"][0]["message"]["content"].strip() data = r.json()
text = data.get("message", {}).get("content", "").strip()
if text.startswith("```"): if text.startswith("```"):
text = re.sub(r"^```\w*\n?", "", text) text = re.sub(r"^```\w*\n?", "", text)
text = re.sub(r"\n?```$", "", text) text = re.sub(r"\n?```$", "", text)