All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- PIPELINE-1407: Changes
Removes the char
'
present in install_requirements. It fixes the issue at build: 'install_requires' must be a string or list of strings containing valid project/version requirement specifiers.
- GlobalFishingWatch/gfw-eng-tasks#44: Adds a fix bug when range didn't include features. Also adds missing test files
- GlobalFishingWatch/gfw-eng-tasks#34: Adds
- Removes commented code
- Pull out dependence on ujson and NewLineJson since no longer used
- Reinstate padding; factor out new hash function so can be used from feature pipeline
- Disable padding during fishing inference
- Fix padding; remove approx_means
- Fix pathname when generaring training paths
- Switch to using blake2b for hashing
- Fix padding bug that generated a lot of fishing regions with bad timestamps
- Add missing input functions
- Fix type causing fail when range was done
- Remove debuggin logging
- Bug fix
- Fix ranges in vessel classification
- More logging
- More debugginf logging
- Change logging to warning to make sure gets through
- Add debugging logging for time range computation
- Use research vessels again
- Stop converting id to int when running inference
- Force zip to return list under py3
- Fix speed problems and memory leaks
- Python 3 compatibility; much directed at working around change in builtin hash
- Simple python 3 fixes; mostly in tests
- Fix print statements
- Separate cargo and tanker in coarse mapping
- Improve metrics computation and add some support for auto generating docs from metadata
- Add model that uses depths so we can test depth inference
- Dont apply synonyms to top level classes
- Automatically convert seismic_vessels to research
- Tweak compute_metrics to keep fine classification table in defined order
- Reinstate seismic vessel as unused class; improvements to metrics
- Fix compute vessel metrics
- Remove seismic vessel and update training and testing to use vessel database
- Debug training data generation and update metric computation.
- Silence yaml warning by switching to safe_load
- Tweak training invokation to try to improve speed
- Switch to computing vessel parameters in terms of MMSI
BREAKING CHANGE, requires pipe-tools and pipe-features 3.0
- #41
- Refactor to use new Tensorflow Dataset and Estimator APIs
- Support more recent tensorflow versions.
- Go back to original random rolling of data during training since tests showed slightly better accuracy.
- Changes to support UVI and MMSI simultaneously.
- Fix way vessel types are upsampled
- Fix vessel metrics to work with vessel_id.
- Correctly stratify data using new classes.
- Generate training data directly from Vessel Database
- Modify training invocation to make tracking runs easier.