meaningless update

README.md

@@ -1,42 +1,42 @@
 # Model Promotion Pipeline Example
 
-A fraud detection pipeline that demonstrates three key `obproject` features working together:
+A fraud detection pipeline demonstrating CI-driven model promotion across branches.
 
 | Feature | Where | What it does |
 |---|---|---|
 | `@project_schedule` | TrainFlow | Cron schedule on main only; other branches deploy without a schedule |
 | `@project_trigger` | EvaluateFlow | Automatically runs when TrainFlow publishes `training_complete` |
-| `promote_assets` | EvaluateFlow | Copies the model asset from a feature branch to main after validation |
+| `promote_assets` | CI workflow | Promotes assets from feature branch to main on PR merge |
 
 ## Pipeline
 
 ```
-TrainFlow (scheduled weekday mornings on main)
-    |
-    | publishes "training_complete" event
-    v
-EvaluateFlow (triggered by event)
-    |
-    | if accuracy >= 0.85 and not on main: promote model to main
-    | trigger_on_finish
-    v
-ReportFlow (prints latest model summary)
+Feature branch:
+  TrainFlow (manual trigger) → EvaluateFlow (pass/fail gate) → ReportFlow
+
+On PR merge:
+  CI promotes assets to main with @candidate → teardown feature branch
+
+Main branch:
+  TrainFlow (weekday 8 AM ET cron) → EvaluateFlow → ReportFlow
 ```
 
 ## Main vs Feature Branches
 
+**On a feature branch:**
+- TrainFlow deploys but has no cron (triggered manually)
+- EvaluateFlow auto-triggers after training, reports pass/fail
+- Models stay on the feature branch until PR merge
+
+**On PR merge:**
+- CI runs `promote_assets()` to copy assets to main with `@candidate` alias
+- CI tears down the feature branch (flows, assets, metadata)
+
 **On main:**
 - TrainFlow runs on a cron (`0 8 * * 1-5` ET)
 - EvaluateFlow auto-triggers after training
-- No promotion happens (already on main)
 - ReportFlow prints the latest model
 
-**On a feature branch:**
-- TrainFlow deploys but has no cron (can be triggered manually)
-- EvaluateFlow auto-triggers after training
-- If the model passes evaluation, it is promoted to main
-- ReportFlow prints the latest model on the feature branch
-
 ## Local Development
 
 ```bash
@@ -52,7 +52,8 @@ METAFLOW_PROFILE=yellow python flows/report/flow.py run
 
 ## Deploying
 
-Push to a GitHub repository with the included workflow. The CI pipeline will:
-1. Deploy all three flows to the platform
-2. On main, TrainFlow gets its cron schedule
-3. On branch deletion/merge, teardown removes branch resources
+Push to GitHub. The CI pipeline handles three lifecycle events:
+
+1. **Push** — deploys flows to the platform (main gets cron, feature branches don't)
+2. **PR merge** — promotes assets to main with `@candidate`, then tears down the feature branch
+3. **Branch delete** — tears down branch resources

flows/evaluate/flow.py

@@ -1,13 +1,13 @@
 """
-Evaluate the latest fraud classifier and promote to main if accurate.
+Evaluate the latest fraud classifier.
 
 Triggered automatically by the training_complete event from TrainFlow.
-On feature branches, promotes the model to main if accuracy exceeds
-the threshold. On main, no promotion is needed.
+Evaluates the model against a held-out test set and reports pass/fail.
+Promotion to main is handled by CI on PR merge, not by this flow.
 """
 
 from metaflow import step, current
-from obproject import ProjectFlow, project_trigger, promote_assets
+from obproject import ProjectFlow, project_trigger
 
 
 ACCURACY_THRESHOLD = 0.85
@@ -57,21 +57,13 @@ def start(self):
 
     @step
     def end(self):
-        branch = self.prj.write_branch
         passed = self.eval_accuracy >= ACCURACY_THRESHOLD
-        print(f"Threshold={ACCURACY_THRESHOLD}  passed={passed}  branch={branch}")
+        print(f"Threshold={ACCURACY_THRESHOLD}  passed={passed}")
 
-        if passed and branch != "main":
-            project = self.prj.project
-            result = promote_assets(
-                project, source=branch, target="main",
-                asset="fraud_classifier", kinds=["models"],
-            )
-            print(f"Promotion result: {result}")
-        elif not passed:
-            print("Model did not meet accuracy threshold; skipping promotion.")
+        if passed:
+            print("Model passed evaluation. Will be promoted to main on PR merge.")
         else:
-            print("Already on main; no promotion needed.")
+            print("Model did not meet accuracy threshold.")
 
 
 if __name__ == "__main__":

models/fraud-classifier/asset_config.toml

@@ -5,3 +5,4 @@ description = "Trained fraud detection model (sklearn)"
 [tags]
 framework = "sklearn"
 domain = "fraud"
+status = "baseline"