v0.3.1

v0.3.1 (2024-03-07)

Fix

• fix: Dockerfile arm build (cce1ebe)

Unknown

• Merge pull request #140 from tattle-made/hotfix

Hotfix (68469b1)

• Merge pull request #139 from aatmanvaidya/fix-dockerfiles

fix: Dockerfile arm build (a031c5d)

v0.3.0

v0.3.0 (2024-03-07)

What's Changed

• Fix ci merge versioning by @duggalsu in #129
• Hotfix by @aatmanvaidya in #130
• feat: workers can report to queue by @aatmanvaidya in #131
• Ci merge versioning by @duggalsu in #132
• ci: Test fix versioning and modify docker images by @duggalsu in #133
• Hotfix by @aatmanvaidya in #134
• rebasing development with main by @aatmanvaidya in #137
• feat: workers to search audio and video files by @aatmanvaidya in #136
• merge development to main by @aatmanvaidya in #138

Full Changelog: v0.2.0...v0.3.0

v0.2.0

v0.2.0 (2024-03-05)

What's Changed

• feat: add video queue worker and fix: modify feluda core by @aatmanvaidya in #90
• fix: operator tests support Media Factory by @aatmanvaidya in #91
• feat: md5 hash operator by @aatmanvaidya in #92
• refactor: moving tests to sub-folders by @aatmanvaidya in #94
• fix: video worker handles disconnection to RabbitMQ by @aatmanvaidya in #95
• ci: fix location for video requirment file by @aatmanvaidya in #100
• fix: video operator locaiton in Dockerfile by @aatmanvaidya in #102
• fix: context manager handles vidoes by @aatmanvaidya in #107
• fix: audio operators uses function from a local folder by @aatmanvaidya in #106
• test: md5 hash supports media factory by @aatmanvaidya in #93
• ci: video worker dockerfile for graviton by @aatmanvaidya in #108
• feat: worker for audio operator by @aatmanvaidya in #109
• test: benchmarking scripts for audiovec operator by @aatmanvaidya in #111
• fix: updating vidvec benchmark scripts by @aatmanvaidya in #112
• refactor: config file for video worker by @aatmanvaidya in #114
• ci: audiovec benchmark graviton dockerfile by @aatmanvaidya in #115
• refactor: merging development to main by @aatmanvaidya in #120
• Fix ci merge workflow by @duggalsu in #121
• ci: Fix secrets access by @duggalsu in #122
• ci: Delete workflow by @duggalsu in #124
• Hotfix by @aatmanvaidya in #125
• refactor: updating dev from main by @aatmanvaidya in #127
• refactor: merge development to main by @aatmanvaidya in #128

Full Changelog: v0.1.0...v0.2.0

v0.1.0

v0.1.0 (2024-03-04)

Chore

• chore: deleting docker-compose-standalone and renaming privacy policy (#89) (bf6fb1e)

• chore: adding .env-template (#88) (c556f8b)

• chore: skipping unmaintained tests (c48b724)

• chore: removing video files (bce9759)

• chore: add DMP issue template (63aa375)

• chore: adding audio file (#73) (52d0fd6)

• chore: not skipping image search test (#57) (9f42558)

• chore: revert to commit 8479c38 (98a76de)

• chore: testing ci (e909448)

• chore: text vector debugging (5097840)

• chore: change batchsize & num_workers to handle memory issue, improve exception logging (7cd003e)

Ci

• ci: Add automated semantic versioning

• Renamed and modified docker push vidvec benchmark yml
• Modified tags and dockerfile name for vidvec worker staging yml
• Added init python file with version number
• Added github workflow on merge on main branch
• Added github workflow on merge on development or hotfix branch (972b6a6)

• ci: audiovec worker staging dockerhub tag change (a79620d)

• ci: audio worker staging github workflow (eec380a)

• ci: workflow dockerfile location change (d0826b8)

• ci: workflow to push video worker to docker hub (594fc6c)

• ci: workflow to push video worker to docker hub (730b808)

• ci: updating vid Dockerfiles (8c32315)

• ci: python version change in video dockerfile (#66) (aafde1c)

• ci: giving access to tests for video vec (#61) (5945004)

Documentation

• docs: adding comments to the config-indexer (7bcbfca)

• docs: tests documentation for YOLO and Tesseract operators (#47) (8eefb63)

• docs: YOLO segementation operator (#46) (1a3da78)

• docs: documentation for tesseract ocr operator (37e3e30)

• docs: add Usage Case Studies (31400a6)

Feature

• feat: feluda store supports audio (#78)

• feat: feluda store supports audio

• fix: delete and refresh for ES

• dhore: profiling audio operator (f6987a6)

• feat: add poc multiprocess test (f43646b)

• feat: audio operator to extract embedding vectors (#59)

• feat: audio emebddings

• chore: deleting music files

• chore: renaming files

• docs: documentation for audio embedding operator

• docs: adding work to be done for the operator (484d5ae)

• feat: c-profiling test for video vec (#60)

• feat: c-profiling test for video vec

• feat: test to find time taken for video vec (247f5db)

• feat: add workflow to push vidvec specific operator to dockerhub (17e0d57)

• feat: operator to detect objects using YOLO (#44)

• feat: operator to detect objects using YOLO

• test file comment main function

• chore: moving ultralytics install to opreator (17b9d10)

• feat: operator to extract text in images using tesseract (#40)

• feat: opreator to detect text in images using tesseract

• chore: adding test images and making test multilingual (edec4a9)

• feat: add license (a44e233)

• feat: update docs (19a9217)

• feat: add NER, add text search, add Not Implemented http response for raw query

• Added an operator to do Named Entity Recognition on non-english text
• Created a test and handler to respond to text queries
• Deferred the raw query search for later. It returns an HTTP response 501 now to denote that it hasnot been implemented (d57fb90)

• feat: add pathways for /index depending on request content type (2f9f11f)

• feat: wip workers ready. testing pending (84f9f33)

• feat: indexing via worker in place (612593e)

• feat: test end to end indexing for text, image and video (af6f144)

• feat: create generator for video vectors and store it in es (0208259)

• feat: index all 3 mediatypes via URL (bcb0237)

• feat: remove stray files (55c39cf)

Fix

• fix: video es test (#87) (2670490)

• fix: url media link for index api test (089412a)

• fix: index and search tests (ffc8efc)

• fix: search as client test (#83) (9855e96)

• fix: graviton supported github runner (1ac59c9)

• fix: based image for graviton (e51ab2f)

• fix: workflow file (8e74e9e)

• fix: video search (#52)

• chore: moving test files to a folder

• fix: video search

• docs: commenting TODO in search.py (af54ac0)

• fix: search api as client (2573490)

• fix: vid_vec_rep_res operator (0efc971)

• fix: index image api (bc1a56a)

• fix: image_vec_rep_resnet operator (7b3b419)

• fix: elastic search test (0326407)

• fix: uncommenting ES_HOST code (144c828)

• fix: server is up and running (29b6100)

• fix: removed trailing comma (a687377)

• fix: add separator between filenames (184022f)

Refactor

• refactor: removing api folder (91c797b)

• refactor: vidvec...

Third Party Models

• This release is to upload and all the third party models that Feluda uses for its operators.

1. PANNs inferece - https://github.com/qiuqiangkong/panns_inference
   PANN is a CNN that is pre-trained on lot of audio files. They have been used for audio tagging and sound event detection. The PANNs have been used to fine-tune several audio pattern recognition tasks, and have outperformed several state-of-the-art systems. Feluda uses PANN's to extract a 2048 dimension vector of any given audio file.
   The pth file of the CNN model will be used in the audio_vec_embedding operator.