fix(embeddings): fall back to default cache dir when FASTEMBED_CACHE_PATH is empty (#3434)

docker-compose.yml injects FASTEMBED_CACHE_PATH=${FASTEMBED_CACHE_PATH:-},
which sets the variable to an empty string when the host has not defined it.
FASTEMBED_CACHE_DIR used os.getenv("FASTEMBED_CACHE_PATH", default), and
os.getenv only returns the default when the variable is ABSENT -- so the empty
value won and FASTEMBED_CACHE_DIR became "". os.makedirs("") then raised
[Errno 2] No such file or directory: '', FastEmbed failed to initialise, and
every vector feature (RAG, semantic memory, tool index) silently degraded on
the default Docker stack.

Treat an empty value like an absent one via `os.getenv(...) or default`.
Add a regression test covering the empty, unset, and explicit cases.

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
TheDragonTail
2026-06-16 03:11:48 +01:00
committed by GitHub
parent 7b09491557
commit 0f966d6b9f
2 changed files with 76 additions and 1 deletions
+7 -1
View File
@@ -57,7 +57,13 @@ MEMORY_VECTORS_DIR = os.path.join(DATA_DIR, "memory_vectors")
# Paths with an intentional dedicated env override, defaulting under DATA_DIR.
MAIL_ATTACHMENTS_DIR = os.getenv("ODYSSEUS_MAIL_ATTACHMENTS_DIR", os.path.join(DATA_DIR, "mail-attachments"))
FASTEMBED_CACHE_DIR = os.getenv("FASTEMBED_CACHE_PATH", os.path.join(DATA_DIR, "fastembed_cache"))
# `or` (not os.getenv's default arg) so a PRESENT-but-EMPTY value falls back to
# the default. docker-compose.yml injects `FASTEMBED_CACHE_PATH=${FASTEMBED_CACHE_PATH:-}`,
# which sets the var to "" when the host hasn't defined it. os.getenv(name, default)
# only returns the default when the var is ABSENT, so the empty string would win →
# os.makedirs("") raises [Errno 2] No such file or directory: '' → FastEmbed fails to
# init and all vector features (RAG, semantic memory, tool index) silently degrade.
FASTEMBED_CACHE_DIR = os.getenv("FASTEMBED_CACHE_PATH") or os.path.join(DATA_DIR, "fastembed_cache")
# Agent tool output limits (single source of truth — imported by tool_execution.py,
# tool_implementations.py, agent_tools.py, and any other module that needs them)