loglabs / mltrace

Coarse-grained lineage and tracing for machine learning pipelines.

https://mltrace.readthedocs.io/

loglabs/mltrace Issues

Materializing historical component input/outputs: Technical Requirements
Updated a year ago
Node shifted LTS from 16 to 18 leading to server build failure
Updated a year ago
React Routing Bug - Clicking back and forward button result in blank page
Closed a year ago
[EPIC] mlflow integration
Closed 2 years ago
Converting component creation to be async
Updated 3 years ago1
SA Warning
Updated 3 years ago
Outputs not tracing
Closed 3 years ago
Clean up examples
Closed 3 years ago
Fix bug in materialized view creation
Closed 3 years ago
Metric Monitoring Storage Layer
Updated 3 years ago
[EPIC] Abstractions for monitoring outputs
Updated 3 years ago
Add functionality to define metrics (i.e. accuracy, precision, recall, f1) to run on the above joined table
Closed 3 years ago
Researching tests for ML
Closed 3 years ago
Can't copy filenames in the trace view
Closed 3 years ago1
Information flow labels
Updated 3 years ago
[Feature idea] materializing history in test methods
Updated 3 years ago
[EPIC] Add functionality to run triggers before and after components are run
Closed 3 years ago
Add ability to copy paste filenames on the UI
Closed 3 years ago
Investigation for predefined component triggers
Closed 3 years ago
Explicitly cast IOPointer keys as strings
Updated 3 years ago
Enhance auto_log ability
Updated 3 years ago
Log STDOUT and STDERR to component runs
Updated 3 years ago
Make sure that inputs and outputs do not have any shared IOPointers
Closed 3 years ago
[EPIC] Store artifacts for users
Closed 3 years ago
Add support or follow the OpenLineage Spec
Updated 3 years ago1
Follow the opentelemetry specifications for Tracing Entities
Updated 3 years ago1
Modify `input_vars` and `output_vars` in decorator to accept non-string variables
Updated 3 years ago
Fix ability to add notes to CR info card in the UI
Updated 3 years ago
Ask user for path to store mltrace artifacts (or implicitly store in home mltrace directory `os.path.join(os.path.expanduser('~'), '.mltrace')`)
Closed 3 years ago
Write functions to load and save artifacts as joblib files (`mltrace.log`, `mltrace.read`)
Closed 3 years ago
fix error message in `get_component_run_information`
Updated 3 years ago
Automatically create component if component does not exist and component run is logged
Closed 3 years ago
allow output ids to have spaces
Updated 3 years ago
Rewrite the `trace` functions to reflect the new `IOPointer` schema.
Closed 3 years ago
[EPIC] Extend `IOPointer`s to store values in addition to keys
Updated 3 years ago1
Extend the `register` decorator and `log_component_run` functions to log KV pairs for I/O. Rewrite `set_dependencies_from_inputs` to reflect the new `IOPointer` schema.
Closed 3 years ago
Modify `store.get_io_pointer` to return the `IOPointer` correctly. Modify `IOPointer` client-facing abstractions but treat the value as null.
Closed 3 years ago
[Feature] Ability to "unflag all"
Closed 3 years ago1
Create and run a DB migration to store values in addition to keys in the `IOPointer`. Change the PK to be (key, value) instead of just key. In the migration, old records should have null values.
Closed 3 years ago
Document windows setup
Updated 3 years ago
Make example of flagging / unflagging outputs
Updated 3 years ago
Modify the UI and CLI to reflect all of the changes from the IOPointer rewrite.
Updated 3 years ago
Make a "forwards" tracing function in the store.
Updated 3 years ago
Ticket-making when there are ML bugs
Updated 3 years ago
Kubeflow integration
Updated 3 years ago
Fix `register`
Closed 3 years ago
Fixing the configuration of DB_SERVER
Updated 3 years ago
[Bug] CR Info card in trace view doesn't show the right notes
Closed 3 years ago
Write documentation for causal analysis
Closed 3 years ago
add `flag` and `unflag` as UI commands
Closed 3 years ago