feat: OllamaClient.identify_speakers() and summarize() for diarization pipeline

This commit is contained in:
2026-04-02 01:03:40 +02:00
parent b8cc8a3b33
commit 9b5b89e159
2 changed files with 103 additions and 0 deletions
+44
View File
@@ -35,3 +35,47 @@ async def test_list_models_returns_list():
client = OllamaClient(base_url="http://localhost:11434")
models = await client.list_models()
assert "gemma3:12b" in models
@pytest.mark.asyncio
async def test_identify_speakers_returns_dict():
import respx, httpx, json
from llm import OllamaClient
client = OllamaClient()
mapping = {"SPEAKER_00": "Thomas", "SPEAKER_01": "Möller"}
transcript_excerpt = "SPEAKER_00: Gut, Herr Möller.\nSPEAKER_01: Danke, Thomas."
with respx.mock:
respx.post("http://localhost:11434/api/generate").mock(
return_value=httpx.Response(200, json={"response": json.dumps(mapping)})
)
result = await client.identify_speakers(transcript_excerpt)
assert result == {"SPEAKER_00": "Thomas", "SPEAKER_01": "Möller"}
@pytest.mark.asyncio
async def test_identify_speakers_returns_empty_on_parse_failure():
import respx, httpx
from llm import OllamaClient
client = OllamaClient()
with respx.mock:
respx.post("http://localhost:11434/api/generate").mock(
return_value=httpx.Response(200, json={"response": "kein json hier"})
)
result = await client.identify_speakers("irgendwas")
assert result == {}
@pytest.mark.asyncio
async def test_summarize_returns_string():
import respx, httpx
from llm import OllamaClient
client = OllamaClient()
with respx.mock:
respx.post("http://localhost:11434/api/generate").mock(
return_value=httpx.Response(200, json={"response": "# Zusammenfassung\n\nKurzer Text."})
)
result = await client.summarize("Thomas: Hallo.\nMöller: Hi.", model="gemma3:12b")
assert "Zusammenfassung" in result