kimi/Timmy-time-dashboard
Archived
1
0
Fork 0
forked from Rockachopa/Timmy-time-dashboard
Code Pull Requests Activity

[claude] Fix pre-existing ruff lint errors blocking git hooks (#1247) (#1248)

Browse Source
This commit is contained in:
Claude (Opus 4.6)
2026-03-23 23:33:37 +00:00
parent fedd164686
commit b5fb6a85cf
23 changed files with 116 additions and 130 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
src/timmy/memory/embeddings.py
View File

@@ -7,10 +7,11 @@ Also includes vector similarity utilities (cosine similarity, keyword overlap).
"""
import hashlib
import json
import logging
import math
import json
import httpx # Import httpx for Ollama API calls
import httpx # Import httpx for Ollama API calls
from config import settings
@@ -20,14 +21,21 @@ logger = logging.getLogger(__name__)
EMBEDDING_MODEL = None
EMBEDDING_DIM = 384 # MiniLM dimension, will be overridden if Ollama model has different dim
class OllamaEmbedder:
"""Mimics SentenceTransformer interface for Ollama."""
def __init__(self, model_name: str, ollama_url: str):
self.model_name = model_name
self.ollama_url = ollama_url
self.dimension = 0 # Will be updated after first call
self.dimension = 0 # Will be updated after first call
def encode(self, sentences: str | list[str], convert_to_numpy: bool = False, normalize_embeddings: bool = True) -> list[list[float]] | list[float]:
def encode(
self,
sentences: str | list[str],
convert_to_numpy: bool = False,
normalize_embeddings: bool = True,
) -> list[list[float]] | list[float]:
"""Generate embeddings using Ollama."""
if isinstance(sentences, str):
sentences = [sentences]
@@ -43,9 +51,9 @@ class OllamaEmbedder:
response.raise_for_status()
embedding = response.json()["embedding"]
if not self.dimension:
self.dimension = len(embedding) # Set dimension on first successful call
self.dimension = len(embedding) # Set dimension on first successful call
global EMBEDDING_DIM
EMBEDDING_DIM = self.dimension # Update global EMBEDDING_DIM
EMBEDDING_DIM = self.dimension # Update global EMBEDDING_DIM
all_embeddings.append(embedding)
except httpx.RequestError as exc:
logger.error("Ollama embeddings request failed: %s", exc)
@@ -59,6 +67,7 @@ class OllamaEmbedder:
return all_embeddings[0]
return all_embeddings
def _get_embedding_model():
"""Lazy-load embedding model, preferring Ollama if configured."""
global EMBEDDING_MODEL
@@ -69,8 +78,13 @@ def _get_embedding_model():
return EMBEDDING_MODEL
if settings.timmy_embedding_backend == "ollama":
logger.info("MemorySystem: Using Ollama for embeddings with model %s", settings.ollama_embedding_model)
EMBEDDING_MODEL = OllamaEmbedder(settings.ollama_embedding_model, settings.normalized_ollama_url)
logger.info(
"MemorySystem: Using Ollama for embeddings with model %s",
settings.ollama_embedding_model,
)
EMBEDDING_MODEL = OllamaEmbedder(
settings.ollama_embedding_model, settings.normalized_ollama_url
)
# We don't know the dimension until after the first call, so keep it default for now.
# It will be updated dynamically in OllamaEmbedder.encode
return EMBEDDING_MODEL
@@ -79,7 +93,7 @@ def _get_embedding_model():
from sentence_transformers import SentenceTransformer
EMBEDDING_MODEL = SentenceTransformer("all-MiniLM-L6-v2")
EMBEDDING_DIM = 384 # Reset to MiniLM dimension
EMBEDDING_DIM = 384 # Reset to MiniLM dimension
logger.info("MemorySystem: Loaded local embedding model (all-MiniLM-L6-v2)")
except ImportError:
logger.warning("MemorySystem: sentence-transformers not installed, using fallback")
@@ -107,13 +121,12 @@ def embed_text(text: str) -> list[float]:
if model and model is not False:
embedding = model.encode(text)
# Ensure it's a list of floats, not numpy array
if hasattr(embedding, 'tolist'):
if hasattr(embedding, "tolist"):
return embedding.tolist()
return embedding
return _simple_hash_embedding(text)
def cosine_similarity(a: list[float], b: list[float]) -> float:
"""Calculate cosine similarity between two vectors."""
dot = sum(x * y for x, y in zip(a, b, strict=False))