Docker + compose packaging, and a home for the artifacts - #272
Merged
Merged
Conversation
…#19, #20) Flavormancer could only run on the machine it was built on. This makes it installable. Dockerfile is deliberately a RUNTIME image. Training needs 32 cores and hours; serving needs neither, so the ~1 GB of trained forests and parquet tables mount at run time instead of baking into a layer that would go stale the moment a head is retrained. Base is slim-bookworm rather than alpine because RDKit ships manylinux wheels that want glibc, and building it on musl is hours of pain for nothing. Runs unprivileged; the health check allows a 180s start period because a cold start genuinely loads 190 heads. docker-compose.yml is Compose and not Kubernetes on purpose — one box, a handful of users, which is the real shape of an on-prem install for a flavour house. Postgres is the pgvector image because the substitution index is a nearest-neighbour search over the 177-dimension profile vector, which is exactly what pgvector exists for (#20). The schema in infra/initdb keeps tox OUT of that vector: safety is not a flavour-match dimension and must not steer "what tastes similar". The interesting part is FLAVORMANCER_HOME. Every artifact path in predict.py and app.py was relative to the working directory, because the systemd deployment has always had code and models sharing one folder. That makes the obvious container mount impossible: any bind reaching the models would also shadow app.py, and the container would start with no application code. I wrote that exact bug into the first draft of the compose file and caught it reviewing the mount. So artifacts now resolve through predict.artifact() under FLAVORMANCER_HOME, which defaults to "." — the existing systemd deployment is byte-for-byte unaffected, and a container can bake the code in while mounting the data. Verified both directions: unset resolves to the working directory, set redirects, full suite passes either way. Docker is not installed on the build box, so the image has NOT been built and run end to end. The compose file parses and the schema is plain SQL, but the first real `docker compose up` should be treated as the test. Called out in the PR rather than implied to be verified. Signed-off-by: Austin L. <86896075+rvnminers-A-and-N@users.noreply.github.com>
rvnminers-A-and-N
added
area:infra
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #19. Closes #20.
Flavormancer could only run on the machine it was built on. This makes it installable.
Dockerfile — a runtime image, not a training one
Training needs 32 cores and hours; serving needs neither. The ~1 GB of trained forests and parquet tables mount at run time instead of baking into a layer that would go stale the moment a head is retrained. Base is
slim-bookwormrather than alpine because RDKit ships manylinux wheels that want glibc. Runs unprivileged; health check allows a 180s start period because a cold start genuinely loads 190 heads.Compose — one box, deliberately
Compose and not Kubernetes: one machine, a handful of users, which is the real shape of an on-prem install for a flavour house. Postgres is the pgvector image because the substitution index is a nearest-neighbour search over the 177-dimension profile vector (#20). The schema keeps tox out of that vector — safety isn't a flavour-match dimension and must not steer "what tastes similar".
The interesting bug, caught in review
Every artifact path in
predict.py/app.pywas working-directory relative, because the systemd deployment has always had code and models in one folder. That makes the obvious mount impossible: any bind reaching the models would also shadowapp.py, and the container would start with no application code.I wrote that exact bug into the first draft of this compose file and caught it reviewing the mount line.
Fix: artifacts resolve through
predict.artifact()underFLAVORMANCER_HOME, defaulting to".".Verified both directions; full suite passes either way.
Not verified
Docker is not installed on the build box, so the image has not been built and run end to end. The compose file parses, the schema is plain SQL, and the path logic is tested — but the first real
docker compose upshould be treated as the test, not as a confirmation. Saying so plainly rather than implying it's proven.Also
CHANGELOG.mdseeded, with v0.2.0 written up and known limits stated (the incoherent heads, the 58 still indicative, Track B not yet running).