v3.9.0

What's Changed

• Removing logging of recieved logs by @jpodivin in #466
• Implementing custom resolver to cover potential security vulnerability by @jpodivin in #470

Full Changelog: v3.8.0...v3.9.0

v3.8.0

What's Changed

• Handling logs as raw text in /analyze endpoints by @Jany26 in #464

Full Changelog: v3.7.0...v3.8.0

v3.7.0

What's Changed

• Minor gitlab fixes by @jpodivin in #461
• Revert "Moving to llama.cpp version b7974" by @jpodivin in #465
• Deployment configuation changes by @jpodivin in #463
• Custom limits for AsyncOpenAIClient by @jpodivin in #467

Full Changelog: v3.6.0...v3.7.0

v3.6.0

What's Changed

• Changing compose file to use vLLM for production config by @jpodivin in #456
• Improving regexes (for consistency with website) by @Jany26 in #458
• Constraining builds and tests to latest Fedora releases by @jpodivin in #460
• Updating huggingface-hub dependency range to >=1.0.0,<2.0.0 by @jpodivin in #459
• Notifying Gitlab about opting out of LD by @Jany26 in #457
• Setting version to 3.6.0 by @jpodivin in #462

Full Changelog: v3.5.0...v3.6.0

v3.5.0

What's Changed

• Pinning inference container to the last known working version by @jpodivin in #449
• Remove PYTHONPATH variable and instead use invocation with interpreter by @jpodivin in #450
• Dependency adjustment and normalization by @jpodivin in #451
• Returning a result for /queue/print endpoint by @Jany26 in #453
• More descriptive logging during LLM call by @Jany26 in #452
• Redacting personal information from logs before analyzing by LLM by @Jany26 in #454
• Setting version to 3.5.0 by @Jany26 in #455

Full Changelog: v3.4.0...v3.5.0

v3.4.0

What's Changed

• Switch to Fedora 43 base image didn't include update of python version in PYTHONPATH override by @jpodivin in #445
• Reworking visibility of certainty ratings by @jpodivin in #446

Full Changelog: v3.3.0...v3.4.0

v3.3.0

What's Changed

• Resolving Caught handled exception, but response already started. when artifacts are missing by @jpodivin in #438
• Adding interface server to dev and prod composes by @jpodivin in #439
• Removing LLAMA_ARG_N_PARALLEL setting from production compose by @jpodivin in #440
• Removing alternative prompts in old format by @jpodivin in #441
• Setting LOGDETECTIVE_PROMPT_TEMPLATES var in the .env file by @jpodivin in #442
• Image update by @jpodivin in #443

Full Changelog: v3.2.0...v3.3.0

v3.2.0

What's Changed

• Flash attention enabling variable, disabled by default by @jpodivin in #434
• Revision of prompts for smaller models by @jpodivin in #433
• Checking merge requests by combination of project id and MR iid by @jpodivin in #435
• Regenerating the lock file by @jpodivin in #436
• Bumping version to 3.2.0 by @jpodivin in #437

Full Changelog: v3.1.0...v3.2.0

v3.1.0

What's Changed

• Moving to prompts defined as Jinja templates by @jpodivin in #430
• Moving from Mistral to Granite model family by @jpodivin in #431
• Rewriting our description field by @jpodivin in #432

Full Changelog: v3.0.0...v3.1.0

v3.0.0

What's Changed

• Simplify reports by removing rendered plots and instead returning JSON data by @jpodivin in #424
• Normalizing CLI options to use dashes by @jpodivin in #428

Full Changelog: v2.13.0...v3.0.0