From 71e67ab65cfe168d77a3b92550bc7189e9a0ac5a Mon Sep 17 00:00:00 2001
From: windyboy <windyboy@gmail.com>
Date: Fri, 13 Feb 2026 09:54:11 +0800
Subject: [PATCH] Add schema versioning to all remaining data files

Adds vlm_schema_version field to all data file formats for schema
evolution and migration tracking. Task #1 already added versioning
to identities.json; this completes the remaining files.

Changes to scanner.py:
- Add "# vlm_schema_version: 1.0" as first comment line in inventory.csv

Changes to commands/analyze.py:
- Add "vlm_schema_version": "1.0" to analysis.json root

Changes to planner.py:
- Add "vlm_schema_version": "1.0" to plan.json root

Changes to executor.py:
- Add "vlm_schema_version": "1.0" to rollback log JSON

Schema versions:
- inventory.csv: 1.0 (in comment header)
- identities.json: 1.0 or 2.0 (set by parse command)
- analysis.json: 1.0
- plan.json: 1.0
- rollback_log.json: 1.0

Future migrations:
- Version field allows detecting old formats
- Can implement auto-migration logic
- Can warn users about version mismatches
- Enables schema evolution without breaking changes

Testing:
- All 449 tests pass
- No test changes needed (backward compatible)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---
 src/vlm/commands/analyze.py | 1 +
 src/vlm/executor.py         | 1 +
 src/vlm/planner.py          | 1 +
 src/vlm/scanner.py          | 1 +
 4 files changed, 4 insertions(+)

diff --git a/src/vlm/commands/analyze.py b/src/vlm/commands/analyze.py
index 54a886a..3c0996f 100644
--- a/src/vlm/commands/analyze.py
+++ b/src/vlm/commands/analyze.py
@@ -97,6 +97,7 @@ def analyze_cmd(
             }
         )
     analysis_data = {
+        "vlm_schema_version": "1.0",
         "metadata": {
             "generated": generation_timestamp,
             "source_identities": str(input),
diff --git a/src/vlm/executor.py b/src/vlm/executor.py
index deeb592..9573d2e 100644
--- a/src/vlm/executor.py
+++ b/src/vlm/executor.py
@@ -365,6 +365,7 @@ class ExecutionEngine:
         """
         # Convert rollback log to JSON-serializable format
         log_data = {
+            "vlm_schema_version": "1.0",
             "log_id": rollback_log.log_id,
             "execution_plan_id": rollback_log.execution_plan_id,
             "executed_at": rollback_log.executed_at.isoformat(),
diff --git a/src/vlm/planner.py b/src/vlm/planner.py
index 04dfe08..d1da521 100644
--- a/src/vlm/planner.py
+++ b/src/vlm/planner.py
@@ -427,6 +427,7 @@ def save_plan(plan: ExecutionPlan, output_path: Path) -> None:
     """
     # Convert ExecutionPlan to dictionary
     plan_dict = {
+        "vlm_schema_version": "1.0",
         "plan_id": plan.plan_id,
         "created_at": plan.created_at.isoformat(),
         "operations": [
diff --git a/src/vlm/scanner.py b/src/vlm/scanner.py
index 985576b..8845c1f 100644
--- a/src/vlm/scanner.py
+++ b/src/vlm/scanner.py
@@ -516,6 +516,7 @@ def save_inventory_csv(files: list[VideoFile], output: Path, library_root: Path)
     
     with open(output, 'w', newline='', encoding='utf-8') as csvfile:
         # Write metadata as comments
+        csvfile.write("# vlm_schema_version: 1.0\n")
         csvfile.write(f"# Generated: {generation_timestamp}\n")
         csvfile.write(f"# Library Root: {library_root}\n")