Skip to content

feat!: Aniparse v2 — probability-based scoring engine - #13

Merged
MeGaNeKoS merged 1 commit into
masterfrom
v2
Feb 22, 2026
Merged

feat!: Aniparse v2 — probability-based scoring engine#13
MeGaNeKoS merged 1 commit into
masterfrom
v2

Conversation

@MeGaNeKoS

Copy link
Copy Markdown
Owner

Summary

  • Complete rewrite with a six-stage pipeline: Tokenize → Identify → Expand → Score → Resolve → Compose
  • Probability-based scoring replaces hardcoded heuristics — confidence accumulates from position, context, keywords, and patterns
  • Structured output with SeriesInfo, sequences, ranges, and alternatives
  • Alternative title support via | separator
  • Folder/path context parsing
  • Extensible keyword system with WordListManager
  • Modernized packaging (pyproject.toml + setuptools-scm for version from git tags)
  • Updated CI workflows (Python 3.7–3.13, modern GitHub Actions)

BREAKING CHANGE: Output format has completely changed. See README for the new schema.

Closes #1, Closes #3

Test plan

Complete rewrite of the parser with a new architecture:
- Six-stage pipeline: Tokenize → Identify → Expand → Score → Resolve → Compose
- Probability-based scoring replaces hardcoded heuristics
- Structured output with SeriesInfo, sequences, ranges, and alternatives
- Alternative title support via pipe separator
- Folder/path context parsing
- Extensible keyword system with WordListManager
- Modernized packaging (pyproject.toml + setuptools-scm)
- Updated CI workflows (Python 3.7–3.13)

Closes #1, Closes #3

BREAKING CHANGE: Output format has changed. See README for the new schema.
@MeGaNeKoS
MeGaNeKoS merged commit b149a3d into master Feb 22, 2026
0 of 28 checks passed
@MeGaNeKoS
MeGaNeKoS deleted the v2 branch February 22, 2026 07:56
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Improve the episode alternative handler (parser.validator) Improve Title and alternative title detection.

1 participant