Add conflict-of-interest check engine; gate stage advancement at Intake

- New fl.conflict.check model: screens petitioner/respondent/party_ids names
  against parties on other open cases (exact partner match + difflib fuzzy
  match at 0.85 threshold); skips folded/closed stages
- Runs automatically as the first action in fl.case.create; logs conflicts to
  chatter with matched-case detail and never silently passes
- fl.case gains conflict_check_passed/conflict_check_id/conflict_check_ids;
  write() blocks advancing stage_id past Intake until the check passes
- Admin-only action_override requires a written justification, stamps user/date,
  and flips conflict_check_passed True with a chatter audit entry
- Add conflict check form/tree/search views, action, Cases sub-menu item,
  case form banner + Run Conflict Check button, and Kanban conflict badge
- ACL entries for fl.conflict.check (admin full, paralegal no-delete)
- Finish Claude migration cleanup in fl_analysis.py (model_used default,
  docstring/help text)
- Add .gitignore for Python artifacts

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

activeblue_familylaw/models/fl_analysis.py

@@ -2,10 +2,7 @@ from odoo import fields, models
 
 
 class FlAnalysis(models.Model):
-    """
-    Phase 5 — Full Ollama AI analysis implementation.
-    Phase 1: Stub with fields required by fl_case computed fields.
-    """
+    """AI case analysis record. Written by the Claude API engine (fl.ai.engine)."""
     _name = 'fl.analysis'
     _description = 'AI Case Analysis Result'
     _order = 'create_date desc'
@@ -19,7 +16,7 @@ class FlAnalysis(models.Model):
     )
     model_used = fields.Char(
         string='AI Model',
-        default='llama3.1'
+        default='claude-sonnet-4-20250514'
     )
 
     # ── Results (referenced by fl_case related fields) ─────────────────────
@@ -67,7 +64,7 @@ class FlAnalysis(models.Model):
     ], string='Case Complexity')
     raw_response = fields.Text(
         string='Raw AI Response',
-        help='Full JSON response from Ollama — for debugging'
+        help='Full JSON response from the Claude API — for debugging'
     )
     error_message = fields.Text(
         string='Error (if analysis failed)'