docs: Add MongoDB data-source and offline-store reference documentation#6351
Conversation
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
…tion Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
jvincent-mongodb
force-pushed
the
master
branch
from
a2d2ca9 to
3eab42a
Compare
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
|
@ntkathole - This is ready for your review/merge. TIA! |
| Use `retrieve_online_documents_v2()` to perform similarity search: | ||
|
|
||
| ```python | ||
| results = FeatureStore.store.retrieve_online_documents_v2( |
There was a problem hiding this comment.
Fine for example, but user copying same block might face issue.
store = FeatureStore(repo_path=".")
results = store.retrieve_online_documents_v2(...)
|
|
||
| ## Key Optimizations | ||
|
|
||
| * **K-collapse**: Multiple FeatureViews that share the same join keys are queried in a single aggregation using `feature_view: {$in: [...]}`, reducing round trips. |
There was a problem hiding this comment.
The actual get_historical_features implementation loops over each proj_name in fv_to_features and issues separate MongoDB aggregation pipelines per feature view.
K-collapse can be enhancements?
There was a problem hiding this comment.
You're right. K-collapse was removed after the complication around projection keys was discovered. Two projections, same feature view. Thank you for pointing this out. I may have to change docstrings too.
|
@jvincent-mongodb Also want to include Data Source and offline store in |
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
|
Hi @ntkathole! Could you please take another look at this? |
|
@jvincent-mongodb Let's update |
Signed-off-by: jvincent-mongodb <jeffrey.vincent@mongodb.com>
|
|
||
| ## 📦 Functionality and Roadmap | ||
|
|
||
| The list below contains the functionality that contributors are planning to develop for Feast. |
There was a problem hiding this comment.
@ntkathole This is a strange line. Isn't the list below functionality that contributors HAVE developed for Feast?
|
|
||
| The MongoDB online store supports [Atlas Vector Search](https://www.mongodb.com/docs/atlas/atlas-vector-search/), enabling similarity search over feature embeddings stored in MongoDB Atlas. This is powered by the `$vectorSearch` aggregation stage and requires MongoDB Atlas (or the `mongodb/mongodb-atlas-local` Docker image for local development). | ||
|
|
||
| See [PR #6344](https://github.com/feast-dev/feast/pull/6344) for full implementation details. |
There was a problem hiding this comment.
Should this line be included? It links back to code..
| provider: local | ||
| online_store: | ||
| type: mongodb | ||
| connection_string: mongodb+srv://<user>:<pass>@cluster.mongodb.net |
There was a problem hiding this comment.
You'll see in CI, but this may need # pragma: allowlist secret too.
| top_k=5, | ||
| ) | ||
|
|
||
| # Each result is a (event_timestamp, entity_key_proto, feature_dict) tuple. |
There was a problem hiding this comment.
The result of FeatureStore's version is OnlineResponse. I'd just remove these lines.
Given the change, I suggest reviewing the comments below. They are about the implementation in MongoDB, which IS what they want to hear about (how it works), just not specific to the function above.
| The MongoDB offline store provides support for reading [MongoDBSource](../data-sources/mongodb.md). | ||
| * Uses a single shared collection with a compound index for all FeatureViews, distinguished by a `feature_view` discriminator field. | ||
| * Entity dataframes can be provided as a Pandas dataframe. The offline store converts entity identifiers into serialized entity keys for efficient lookup against the collection. | ||
|
|
There was a problem hiding this comment.
Skip the description.
4 participants
What this PR does / why we need it:
Adds reference documentation for the MongoDB contrib data source and offline store integration. This fills the documentation gap for the MongoDB offline store that was added in #6138.
Changes:
docs/reference/data-sources/mongodb.md— documentsMongoDBSource, configuration, supported typesdocs/reference/offline-stores/mongodb.md— documentsMongoDBOfflineStore, data model, retrieval semantics, materialization, and known limitationsdocs/SUMMARY.mdanddocs/reference/data-sources/README.mdto include MongoDB in the navigationWhich issue(s) this PR fixes:
Follows up on #6138 (feat: MongoDB offline store)
Checks
git commit -s)Testing Strategy
This is a docs-only change (no code changes).
Misc
Documentation covers:
feature_historycollection design)