- 8f6c8ef feat: vanilla ae with independent channels (#392)
- 1239acd Multichannel config (#390)
- 56a6ac1 feat: Support Python 3.12; upgrade pynumaflow to 0.7.1 (#383)
- c61f59d fix: release new patch version (#385)
- 4122625 fix: make pivot as optional (#384)
- 309f16a Added multichannel autoencoder with test cases (#382)
- Avik Basu
- Kushal Batra
- Miroslav Boussarov
- 1b6511b feat: Support Python 3.12; upgrade pynumaflow to 0.7.1
- Avik Basu
- 56a6ac1 feat: Support Python 3.12; upgrade pynumaflow to 0.7.1 (#383)
- Avik Basu
- Kushal Batra
- 4122625 fix: make pivot as optional (#384)
- 309f16a Added multichannel autoencoder with test cases (#382)
- 934d8cf Feat/vanilla ae refactor (#379)
- Avik Basu
- Kushal Batra
- Miroslav Boussarov
- 6f46250 fix: filters
- 87c16da fix: filters
- bbcfd47 fix: threshold and filters
- 984650b feat: agg from conf for multi pivot (#378)
- 402aabf Fix: Changing pivot to pivot_table to support aggregation (#376)
- 141e9a0 Postproc support for None (#375)
- e1ae3ee feat: multi column pivot for druid connector (#374)
- 98c5766 Static filter (#373)
- f9ebfaf fix: add threshold in metadata
- cbf4dbd try: replicate og vanilla ae
- 8b2e939 feat: enable feature transforms
- eed9ad4 fix: adjust factor
- 9ad6d8d feat: fallback to stddev if threshold is too low
- 5593be3 fix: C contiguous error for exp mov avg tx
- 8321914 correct struct log for ack_insuf_data
- b36356c fix: percentile scaler, exp mov avg, sigmoid norm (#369)
- f7bd30b feat: add PercentileScaler
- Avik Basu
- Gulshan Bhatia
- Kushal Batra
- Nandita Koppisetty
- s0nicboOm
- Avik Basu
- Saisharath Reddy Kondakindi
- 8bba9ca Optimize UDF logs (#370)
- Gulshan Bhatia
- 984650b feat: agg from conf for multi pivot (#378)
- Nandita Koppisetty
- 402aabf Fix: Changing pivot to pivot_table to support aggregation (#376)
- Nandita Koppisetty
- 141e9a0 Postproc support for None (#375)
- e1ae3ee feat: multi column pivot for druid connector (#374)
- Kushal Batra
- Nandita Koppisetty
- 98c5766 Static filter (#373)
- Kushal Batra
- f9ebfaf fix: add threshold in metadata
- s0nicboOm
- Avik Basu
- eed9ad4 fix: adjust factor
- Avik Basu
- 9ad6d8d feat: fallback to stddev if threshold is too low
- Avik Basu
- 5593be3 fix: C contiguous error for exp mov avg tx
- Avik Basu
- 8321914 correct struct log for ack_insuf_data
- Gulshan Bhatia
- b36356c fix: percentile scaler, exp mov avg, sigmoid norm (#369)
- Avik Basu
- Avik Basu
- Gulshan Bhatia
- Gulshan Bhatia
- Saisharath Reddy Kondakindi
- ad20add fix: remove print (#363)
- 61f6598 feat: initial support for flattened vector in backtest (#361)
- 0cbaa52 feat: introduce stride in dataset (#360)
- db2cc4f feat: support for train transformers (#354)
- 377ada2 feat: generate static scores for backtesting (#355)
- Avik Basu
- Kushal Batra
- 8dee66b feat!: support static threshold score adjustment (#350)
- Avik Basu
- 4e93088 bump dev version
- 81971dc fix: static threshold
- db943a6 fix: conditional fwd tags
- c0225a9 refactor
- 8f8a532 feat: score adjustment with joined trainer vtx
- ac0eda1 more tests
- 6623e2b feat: trainer join vertex for preprocess & inference
- fb97b3a feat: allow chunked druid fetch (#349)
- d8bda0e feat: Multivariate Anomaly Detection (#343)
- d0dcb47 feat: score adjuster using static threshold
- Avik Basu
- b41a7f7 fix: comments
- 780f2e2 fix merge issues
- 9de1e83 fix: backtest
- c418b5a tmpfix: numalogic opex tags for query filter
- bc6d08c feat: Score config in postprocess
- a913545 fix: get item for final score
- cc681a3 tmp: testing per feature score
- 634b2cf fix: factory class
- 36db5df tests: median thresh tests
- 777b5e2 fix: import
- a075761 feat: add percentile thresholding
- fbf171f fix: pipeline id keyerr in tools.py
- add642c fix: pipeline id key error
- dc35c80 try: better docker workflow
- faf1e32 fix: default ml pipe id in preproc
- 03faf8f fix: vae nsamples
- 16c02f1 feat: add beta parameter for disentanglement
- 5ec5883 feat: difference transform
- 46363d3 feat: output threshold + final scores
- 7f9317c feat: support loading nl conf
- 3035407 feat: multivariate backtesting
- c0596d7 add more logs
- 3ca8c15 tmp: try exp mov average
- 43c1ec8 feat: add transforms and robust thresholding
- 19ddc8e Demo branch (#335)
- Avik Basu
- Kushal Batra
- 4dbde41 Fix: Get metrics from _conf.metrics on Trainer to avoid issue with Flattening the matrix (#346)
- shashank10456
- 824a55d fix: add FlattenVector transformer (#344)
- 146ec00 fix: add unified conf (#342)
- 870f263 fix: take mean before calculating the thresholds (#340)
- dbb510f fix: add max value map for cliping the value (#339)
- 80ef431 "Source" tag for metrics (#338)
- b292553 Add pl conf (#336)
- 28fa28f Metrics (#322)
- Kushal Batra
- dfc383a feat: support both base conf and app conf (#328)
- 8b7f45f feat!: support full multivariate prometheus fetching (#325)
- Avik Basu
- fb80cb6 fix: inf filling
- Avik Basu
- 5aa4c13 fix: use ckeys aligning with config in pre, post and inference vtx
- Avik Basu
- b761c94 fix: composite key unzip
- Avik Basu
- 9d1b999 fix: send conf keys instead of request keys to trainer
- Avik Basu
- 14e1a66 fix: mock method
- 8369510 feat!: separate Prom trainer and Druid trainer
- 7be6ddc add more tests
- c783ba8 feat!: support mv query in fetch() method
- Avik Basu
- 870f263 fix: take mean before calculating the thresholds (#340)
- dbb510f fix: add max value map for cliping the value (#339)
- 80ef431 "Source" tag for metrics (#338)
- Kushal Batra
- 146ec00 fix: add unified conf (#342)
- 870f263 fix: take mean before calculating the thresholds (#340)
- dbb510f fix: add max value map for cliping the value (#339)
- 80ef431 "Source" tag for metrics (#338)
- b292553 Add pl conf (#336)
- 28fa28f Metrics (#322)
- dfc383a feat: support both base conf and app conf (#328)
- 8b7f45f feat!: support full multivariate prometheus fetching (#325)
- c967f20 feat: support multivar prom (#317)
- 46cdfcd add retrain logic for insufficient data (#321)
- 0bbb53d doc: Update to fix examples ipynb files (#318)
- 61f4575 chore!: unify and refactor trainer (#315)
- Avik Basu
- Haripriya
- Kushal Batra
- 9787b2d update version
- a12948f chore!: unify and refactor trainer
- 509e38a feat: support multivariate threshold output (#314)
- Avik Basu
- 64c2e95 add: druidfetcher support for different configId (#307)
- Kushal Batra
- f25f49a feat: add jitter (#305)
- Kushal Batra
- d7f9605 feat: update druid query context (#304)
- shrivardhan
- a0e0ad0 fix: docker extra args error (#302)
- Avik Basu
- Kushal Batra
- shrivardhan
- fdec237 fix: trainer bug (#297)
- Kushal Batra
- d249942 fix druid connector with tests (#296)
- shrivardhan
- 98e376a fix: udf server start error (#294)
- Avik Basu
- f55312b fix: pydruid version update (#293)
- Kushal Batra
- 21f85f9 try : wfl (#290)
- Kushal Batra
- 46cdfcd add retrain logic for insufficient data (#321)
- 0bbb53d doc: Update to fix examples ipynb files (#318)
- 61f4575 chore!: unify and refactor trainer (#315)
- 509e38a feat: support multivariate threshold output (#314)
- 64c2e95 add: druidfetcher support for different configId (#307)
- f25f49a feat: add jitter (#305)
- d7f9605 feat: update druid query context (#304)
- a0e0ad0 fix: docker extra args error (#302)
- dfab26e refactor druid connector (#301)
- 2973dd2 feat: add dedup logic (#299)
- fdec237 fix: trainer bug (#297)
- d249942 fix druid connector with tests (#296)
- 98e376a fix: udf server start error (#294)
- f55312b fix: pydruid version update (#293)
- 21f85f9 try : wfl (#290)
- 3e89ccd fix: allow full df scores in backtest (#288)
- bc1c627 feat: add multiple save for redis registry (#281)
- a364721 feat: initial version of backtest tool (#282)
- 0cdc257 feat: use well-defined dimensions instead of strings (#284)
- de8930a feat!: numalogic udfs (#271)
- c62c902 doc: update quick-start.md (#246)
- 52a65c0 fix: improve metadata serialization (#244)
- 8482c01 perf: improve serialization performance (#243)
- 76cac48 fix: tensor dimension swap instead of view change (#240)
- 2dfd84c feat: convolutional vae for multivariate time series (#237)
- dc18442 feat: Multivariate threshold using Mahalanobis distance (#234)
- 466681b feat: add thread safety to local cache (#224)
- Avik Basu
- Haripriya
- Jason Zesheng Chen
- Kushal Batra
- shrivardhan
- adebf98 feat: add thread safety to local cache (#224)
- Avik Basu
- b21e246 fix (RedisRegistry): avoid overwriting cache with the same key during load (#223)
- 0b0daed feat!: add dynamodb registry (#220)
- dacae49 fix: dataset slicing (#222)
- 7778ae8 fix: update production key to latest key (#221)
- 4831180 examples: pipeline using block (#216)
- 203c100 Upgraded python support for 3.11 (#211)
- 9ebc32f feat!: introduce numalogic blocks (#206)
- 4058811 fix: redis logging (#209)
- b93bce5 fix: removing logger level setting (#205)
- 6727296 fix: add caching logs (#203)
- 4411aa7 examples: update with new numalogic and pynumaflow (#202)
- 2298be4 chore!: refactor preproc and postproc into transforms module (#201)
- 9e838ef fix: Sparse AE for vanilla and conv (#199)
- 3193159 fix: registry test_case (#197)
- efa1df7 chore!: Auto detect instance type while mlflow model save (#190)
- 4a1effd feat: add redis caching (#179)
- 0f7e6e0 fix: allow import from Registryconfig with optional dependencies (#180)
- 6c21e95 fix: stale check; conf lazy imports (#178)
- f1909a8 feat: redis registry (#170)
- 794ddc6 feat: local memory artifact cache (#165)
- 03514d6 chore!: drop support for python 3.8 (#164)
- 73bbad2 feat: first benchmarking using KPI anomaly data (#163)
- ed40681 feat: support weight decay in optimizers (#161)
- cae88b3 chore!: use torch and lightning 2.0 (#159)
- Avik Basu
- Kushal Batra
- Miroslav Boussarov
- Tarun Chawla
- 79987db release: v0.4.1 (#218)
- Avik Basu
- b134fa2 fix: redis logging (#209)
- Kushal Batra
- bd050c9 fix: removing logger level setting (#205)
- Kushal Batra
- 1f5f458 fix: add caching logs (#203)
- ce93191 examples: update with new numalogic and pynumaflow (#202)
- fd169cf chore!: refactor preproc and postproc into transforms module (#201)
- a2b00c1 fix: Sparse AE for vanilla and conv (#199)
- 5e69f5f fix: registry test_case (#197)
- Avik Basu
- Kushal Batra
- 75aea49 chore!: Auto detect instance type while mlflow model save (#190)
- e884b90 feat: add redis caching (#179)
- 7834730 fix: allow import from Registryconfig with optional dependencies (#180)
- 25a16f2 fix: stale check; conf lazy imports (#178)
- Avik Basu
- Kushal Batra
- 54cffe3 fix: optional dependency imports
- Avik Basu
- 8086db1 fix: stale check; conf lazy imports (#178)
- b664e49 Prerelease 0.4 (#173)
- 85fb527 chore!: use torch and lightning 2.0
- Avik Basu
- 3160c2b feat: exponential moving average postprocessing (#156)
- 9de8e4c fix: validation loss not being logged (#155)
- Avik Basu
- b61ac1f fix: Tanhscaler nan output for constant feature (#153)
- 69006eb Update CODEOWNERS (#151)
- 6b38465 feat: more generic convolutional ae (#149)
- Avik Basu
- Vigith Maurice
- cb5509a add: anomaly sign and return labels for anomalies generated (#146)
- b6f63ef fix: latest model calling (#145)
- 6ee3446 fix: transition (#144)
- Kushal Batra
- ea01b44 feat: Sigmoid threshold (#141)
- Avik Basu
- ec8401d feat: tanh preprocessing (#139)
- Avik Basu
- 4eae629 fix!: consistency with threshold methods (#138)
- 2ac1c2f feat: static threshold estimator (#136)
- d3488c9 feat: initial config schema (#135)
- Avik Basu
- 2923718 fix: unbatch error on certain cases (#131)
- 232302e fix: pin protobuf to v3.20 for pytorch-lightning (#130)
- dca9a7a feat!: merge to release v0.3 (#119)
- df20591 fix: sklearn base import for scikit learn v1.2 (#112)
- d2c6293 fix: change example pipeline mlflow port (#96)
- f9c74c9 fix: allow only patch updates in torch version due to cuda build errors on mac (#90)
- 0a5fcdf fix_readme: mention namespace name when applying the pipeline (#88)
- Avik Basu
- Kushal Batra
- 88d26ec feat!: convert AE variants to lightning modules (#110)
- ed94615 fix: fix and clean mlflow test cases (#109)
- Avik Basu
- Kushal Batra
- 78cf5b4 fix: fix pipeline for 0.3 release (#106)
- 709553f fix: fix mlflow test cases (#98)
- 701812e feat!: disentangle threshold selection from the main model (#89)
- Avik Basu
- Kushal Batra
- dca9a7a feat!: merge to release v0.3 (#119)
- Avik Basu
- c00bb14 fix: Upgrade torch to 1.13.1 (#128)
- Avik Basu
- df20591 fix: sklearn base import for scikit learn v1.2 (#112)
- Avik Basu
- 7d5075d remove mlflow full
- 05d6071 fix: have mlflow-server as an optional extra
- 3c5bc83 fix: lock file
- d2c6293 fix: change example pipeline mlflow port (#96)
- Avik Basu
- Kushal Batra
- f9c74c9 fix: allow only patch updates in torch version due to cuda build errors on mac (#90)
- 0a5fcdf fix_readme: mention namespace name when applying the pipeline (#88)
- 537fae5 fix: AutoencoderPipeline logged loss mean (#55)
- 22f8e5d chore!: make mlflow as an optional dependency (#47)
- Avik Basu
- Kushal Batra
- diego-ponce
- 5703d1b fix: update readme with optional mlflow dependency
- 702f3b4 fix: install extras in workflows
- 8969e80 chore!: make mlflow as an optional dependency
- 9cc97cb feat: resume training parameter (#40)
- Avik Basu
- Kushal Batra
- f0b0d3c fix: loading secondary artifacts (#16)
- 0c506bd chore (#19)
- c845d09 Update README.md
- e8be7c5 fix: pypi auto publish workflow (#18)
- fa22031 [Chore] Update README (#17)
- Avik Basu
- Kushal Batra
- Saradhi Sreegiriraju
- Vigith Maurice
- amitkalamkar
- 6847211 workflows: add pypi publish and auto release generation (#14)
- 040584f feat: adding feature for retaining fixed number of stale model (#13)
- Avik Basu
- Kushal Batra
- a5ef072 feat: Add support for storing preproc artifacts (scondary artifact) i… (#11)
- Kushal Batra
- 5314070 feat: add transformers model (#8)
- Kushal Batra