Граф коммитов

972 Коммитов

Автор SHA1 Сообщение Дата
Anurag Aggarwal 656d6e844b Remove bugs_retrieval image and use the base image instead in its place (#691)
* Fixes #633
2019-07-12 14:17:41 +02:00
Marco Castelluccio 98131301ef Clarify unassigned issues rules 2019-07-12 11:54:51 +02:00
Marco Castelluccio e12f4cf040 Download bugs DB when CommitModel's bug_data is set to True
Fixes #690
2019-07-12 10:18:55 +02:00
Marco Castelluccio 2096b9ecef Version 0.0.66 2019-07-12 09:26:41 +02:00
cklyyung 114dae4eb9 Replace deprecated download_bugs_between with get_ids_between (#689) 2019-07-12 09:25:35 +02:00
Marco Castelluccio 09741b82df Remove tests for the old 'title' feature 2019-07-11 20:55:19 +02:00
Marco Castelluccio 84bd06021e Remove other references of the old 'title' feature 2019-07-11 20:54:02 +02:00
Marco Castelluccio d9cdcdc238 Enable feature importance calculation for the defect/enhancement/task model 2019-07-11 20:44:07 +02:00
Marco Castelluccio ca0014450f Use layered violin summary plot only for binary classification, use default plot otherwise 2019-07-11 20:42:56 +02:00
Marco Castelluccio 4d7ebe57a4 Remove unneeded custom 'title' feature
It's a relic from the past, it was True when the title contained the text 'fail'
2019-07-11 18:03:01 +02:00
Harshit chittora d5154c837d Use some bug features in the Backout model (#615)
Fixes #577.
2019-07-11 15:42:02 +02:00
Marco d426d5e66e
Revert "Revert "Remove pip from the commit retrieval image (#676)" (#686)" (#687)
This reverts commit e84c248e19.
2019-07-11 02:02:04 +02:00
Marco e84c248e19
Revert "Remove pip from the commit retrieval image (#676)" (#686)
This reverts commit e7ed31940c.
2019-07-10 16:52:04 +02:00
Marco Castelluccio 9a53baf7ec Upload importances.json files as an artifact 2019-07-10 16:39:48 +02:00
Marco Castelluccio 17b027c767 Enable feature importance calculation at training time for the regressor model 2019-07-10 16:25:38 +02:00
Marco Castelluccio 803003a0d0 Ignore very infrequent words in commit descriptions 2019-07-10 16:25:12 +02:00
Marco Castelluccio 66c496bb25 Don't use 'files' as feature of the regressor model
There are too many, shap calculation runs out of memory if we consider them.
We should add a minimum_frequency parameter so we only return the most often modified files.
2019-07-10 16:17:45 +02:00
Marco Castelluccio 7179cdefc4 Use layered violin summary plot, as it's more readable especially for categorical features 2019-07-10 14:58:57 +02:00
Marco Castelluccio f2c8ce4313 Convert training set to array before generating a summary plot
shap does not support sparse matrices
2019-07-10 14:58:12 +02:00
Marco Castelluccio d7b7ccee74 Store most important features in a JSON file too 2019-07-10 14:57:16 +02:00
Marco Castelluccio bcde3415f4 Version 0.0.65 2019-07-09 21:32:36 +02:00
Marco Castelluccio 2bc8d5c276 Support combined text feature too 2019-07-09 21:23:53 +02:00
Marco Castelluccio 481917afac Use human readable feature names when classifying too 2019-07-09 21:22:37 +02:00
Davide Monfrecola e7ed31940c Remove pip from the commit retrieval image (#676)
Fixes #647
2019-07-09 18:32:57 +02:00
Marco Castelluccio 5c7af453ac Version 0.0.64 2019-07-09 18:14:55 +02:00
Marco Castelluccio be6b652613 Fix typo in reviewers feature name 2019-07-09 18:14:40 +02:00
Marco Castelluccio 47e6a15425 Version 0.0.63 2019-07-09 17:38:00 +02:00
Marco Castelluccio 261bff6ab0 Rename public/probs,json artifact to public/probs.json 2019-07-09 14:41:03 +02:00
Ayush Shridhar f5bd2ea5ad Add cut off to word2vec similarity (#683) 2019-07-09 12:11:50 +02:00
Marco Castelluccio 9f4262bbfe For feature importance, show human readable feature names
Fixes #672
2019-07-09 11:16:18 +02:00
Ayush Shridhar b456bd3b2d Ignore out-of-vucabulary words in word2vec similarity (#680) 2019-07-08 15:22:21 +02:00
Ayush Shridhar bc6467da41 Add word2vec similarity option to evaluation script (#678) 2019-07-08 12:58:49 +02:00
Ayush Shridhar ad2546e124 Add word2vec with WMD distance similarity (#666) 2019-07-08 12:24:13 +02:00
Marco Castelluccio d102ccfa54 Version 0.0.62 2019-07-05 15:41:21 +02:00
Marco Castelluccio e98c6054cd Add more defect/enhancement/task labels gathered from changed made by users on Bugzilla 2019-07-05 15:41:13 +02:00
Boris Feld e7add98563 Update task-boot to 0.1.9 (#675) 2019-07-05 15:36:16 +02:00
Marco Castelluccio ec45dd930d Move more features to the 'historical' case 2019-07-05 13:59:23 +02:00
pyup.io bot 5210d0b764 Update nltk from 3.4.3 to 3.4.4 (#673) 2019-07-05 12:12:43 +02:00
Marco Castelluccio 608d7dc4a3 Remove 'reserved-for-beginners' rule 2019-07-05 11:39:24 +02:00
Marco Castelluccio f4636ed7fe Only use top features in the force_plot, or it can become unreadable 2019-07-04 18:59:52 +02:00
Boris Feld 2997d222b0 Rename the suggestion field into class (#670)
Following this review comment:
https://github.com/mozilla/relman-auto-nag/pull/733/files#r299047812
2019-07-04 12:49:33 +02:00
Marco Castelluccio d08e062c68 Improve triggerSchema of the hooks and test it 2019-07-03 17:26:01 +02:00
Marco Castelluccio aac6326ba5 Index classify-patch task in the Taskcluster index by diff ID
Fixes #668
2019-07-03 15:46:35 +02:00
Marco Castelluccio 28cb2b291e Insert the Phabricator diff ID in the command directly rather than in the env 2019-07-03 15:46:21 +02:00
Marco Castelluccio 3eb3a90830 Don't calculate feature importance, as shap doesn't support multilabel models 2019-07-03 14:34:51 +02:00
Marco Castelluccio 5e90044483 Don't use a sampler, as imbalance-learn doesn't support multilabel models 2019-07-03 14:34:23 +02:00
Marco Castelluccio 7702dce48c Use upper-case for constants in bug_type model 2019-07-03 14:33:41 +02:00
Marco Castelluccio 3315532904 Use mozilla/bugbug-commit-retrieval image for the classify-patch hook
Since we need to use Mercurial
2019-07-03 10:40:02 +02:00
Marco Castelluccio 9de3aa70f2 Add TC_SECRET_ID to the env of the classify-patch hook 2019-07-03 10:39:37 +02:00
Marco Castelluccio 7baa09f7ec Version 0.0.61 2019-07-02 21:39:37 +02:00