Git sync → change parsing → semantic mapping → LLM intent inference → knowledge output.

Five phases auto-flow, progressively transforming Git history into structured business evolution knowledge.

1)    Git sync: auto-sync Git repository, fetching latest commit history and change data.

2)    Change parsing: parse file changes in each commit, extracting code-level change information.

3)    Semantic mapping: map code changes to model/field-level semantic changes.

4)    LLM intent inference: LLM infers business intent of each commit, transforming technical changes to business knowledge.

5)    Knowledge output: structured business evolution knowledge for migration engines and AI coding assistants.

File changes in each commit mapped to model/field-level semantic changes.

commit.semantic.change records module/model/field-level changes, giving code changes business semantics.

1)    Module-level changes: identify modules involved, recording additions, deletions, modifications.

2)    Model-level changes: identify models involved, recording additions, deletions, modifications.

3)    Field-level changes: identify fields involved, recording additions, deletions, type changes.

4)    Method-level changes: identify methods involved, recording additions, deletions, signature changes.

5)    Change correlation: correlate changes at different levels, forming complete change semantic graph.

LLM infers business intent of each commit, transforming technical changes into business knowledge.

Not simple change descriptions, but deep understanding of the business motivation behind changes.

1)    Business intent identification: LLM analyzes changes, infers intent (new feature, bug fix, optimization).

2)    Change impact assessment: evaluates impact on business processes and data models.

3)    Change reason analysis: combines commit info and changes to analyze reasons and context.

4)    Change classification: classifies changes by business dimension (functional, data, process).

Structured business evolution knowledge for migration engines and AI coding assistants.

Knowledge is not stored idle, but served as a service for other systems to consume.

1)    Migration engine utilization: migration engine uses evolution knowledge for version upgrades.

2)    AI coding assistant utilization: AI coding assistant uses knowledge for more accurate suggestions.

3)    Knowledge query API: API interface supporting queries by module, model, time dimensions.

4)    Knowledge visualization: visual interface for intuitive project evolution history.