andgineer / lexiflux / 21141886476 / 1

78%

main: 78%

push

github

Version v2.8.1
 use pagesmith that correctly split HTML with namespaced attribs
 optimized SQLite for parallelization
 special processing for retired models
 make test_detect_language less flaky
 solved the rest of type checker warnings
 suppressed stacked generic decorators issue in urls.py
 Django related type hints
 fixed type hints in smart_login_required add get_custom_user() to all methods
 fixed type hints in smart_login_required add get_custom_user() to all methods
 accept English for short text and low confidence
 do not return detected language as None
 useg epub get_context instead of get_body_context
 fasttext-numpy2 to be compatible with numpy 2
 fix AbstractBaseUser type warnings no Django 6 for Python 3.11
 pyrefly CustomUser fix
 pyrefly
 gemini 3
 gemini 3
 combined AI article for convenience made test to use current models list so it won't break on each models list update
 gemini 2.5 flash as default
 LLM models profiled, selected the best for the lexiflux.
 resume + running total cost
 LLM profiler
 faster LLM openAI model in defaults better LLM prompts
 fixed _extract_title for broken HTML
 deduplicate HTTP Django log linter
 LLM models updated
 sidebar tabs on one line on mobile to save vertical space
 shoter button caption
 no middle layout
 tailscale-compatible
 adaptive layout
 separate mobile layout
 reader keyword
 https://github.com/andgineer/lexiflux/issues/78 koyeb deploy docs
 https://github.com/andgineer/lexiflux/issues/78 koyeb deploy workflow
 https://github.com/andgineer/lexiflux/issues/78 WhiteNoiseMiddleware to simplify static serv in non debug env
 https://github.com/andgineer/lexiflux/issues/78 tests django settings
 https://github.com/andgineer/lexiflux/issues/78 deploy environments
 agents file

126 of 229 branches covered (55.02%)

Branch coverage included in aggregate %.

4006 of 5075 relevant lines covered (78.94%)

1.44 hits per line