Merge pull request #713 from mozilla/philimon/choose_split

Philimon/l10n_choose_split

Author: philimon-reset

.github/workflows/l10n-write-mappings.yml

@@ -0,0 +1,31 @@
+---
+name: L10N output mapping splits
+
+run-name: ${{ github.actor }} is writing L10N mapping splits
+on:
+  schedule:
+    - cron: "0 0 * * 1"
+
+permissions:
+  contents: "write"
+
+jobs:
+  Select-Channels:
+    runs-on: ubuntu-latest
+    outputs:
+      channels: ${{ steps.splits.outputs.channels }}
+    steps:
+      - name: Create app token
+        uses: actions/create-github-app-token@v1
+        id: app-token
+        with:
+          app-id: ${{ secrets.BOT_CLIENT_ID }}
+          private-key: ${{ secrets.BOT_PRIVATE_KEY }}
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          token: ${{ steps.app-token.outputs.token }}
+      - name: Select test channels
+        id: splits
+        run: |
+          python3 choose_l10n_beta_split.py

choose_l10n_beta_split.py

@@ -0,0 +1,38 @@
+import json
+import logging
+import sys
+from collections import defaultdict
+
+from choose_l10n_ci_set import valid_l10n_mappings
+
+
+def distribute_mappings_evenly(mappings):
+    """
+    Distribute the selected mappings into 3 splits and return the container.
+
+    Args:
+        mappings (dict): A dictionary of mappings, where the keys are sites and the values are sets of regions.
+    """
+    if not mappings:
+        return {}
+    # sort the mappings by the length of the regions per site
+    mappings = dict(sorted(mappings.items(), key=lambda val: len(val[1]), reverse=True))
+    # place the mappings into 3 containers evenly according to the load
+    loads = [0, 0, 0]
+    balanced_splits = [defaultdict(list) for _ in range(3)]
+    for site, regions in mappings.items():
+        min_idx = loads.index(min(loads))
+        balanced_splits[min_idx][site] = list(regions)
+        loads[min_idx] += len(regions)
+    return balanced_splits
+
+
+if __name__ == "__main__":
+    l10n_mappings = valid_l10n_mappings()
+    all_splits = distribute_mappings_evenly(l10n_mappings)
+    if not all_splits:
+        logging.warning("No valid l10n mappings")
+        sys.exit(1)
+    for idx, split in enumerate(all_splits, start=1):
+        with open(f"l10n_CM/beta_run_splits/l10n_split_{idx}.json", "w") as f:
+            json.dump(split, f)

choose_l10n_ci_set.py

@@ -34,35 +34,6 @@ def valid_l10n_mappings():
     return mapping
 
 
-def distribute_mappings_evenly(mappings, version):
-    """
-    Distribute the selected mappings if its a reportable run.
-
-    Args:
-        mappings (dict): A dictionary of mappings, where the keys are sites and the values are sets of regions.
-        version (int): The beta_version of the beta.
-    """
-    if not mappings:
-        return {}
-    if os.environ.get("TESTRAIL_REPORT"):
-        # sort the mappings by the length of the regions per site
-        mappings = dict(
-            sorted(mappings.items(), key=lambda val: len(val[1]), reverse=True)
-        )
-        # place the mappings into 3 containers evenly according to the load
-        loads = [0, 0, 0]
-        containers = [defaultdict(set) for _ in range(3)]
-        for key, value in mappings.items():
-            min_idx = loads.index(min(loads))
-            containers[min_idx][key] = value
-            loads[min_idx] += len(value)
-        # get container index according to beta beta_version
-        run_idx = version % 3
-        return containers[run_idx]
-    else:
-        return mappings
-
-
 def process_changed_file(f, selected_mappings):
     """
     process the changed file to add the site/region mappings.
@@ -108,6 +79,21 @@ def save_mappings(selected_container):
         f.writelines(current_running_mappings)
 
 
+def select_l10n_mappings(beta_version):
+    """
+    Select the correct l10n mappings.
+
+    Args:
+        beta_version: the current beta version.
+    """
+    beta_split = (beta_version % 3) + 1
+    if os.path.exists(f"l10n_CM/beta_run_splits/l10n_split_{beta_split}.json"):
+        with open(f"l10n_CM/beta_run_splits/l10n_split_{beta_split}.json", "r") as f:
+            return json.load(f)
+    else:
+        return valid_l10n_mappings()
+
+
 if __name__ == "__main__":
     if os.path.exists(".env"):
         with open(".env") as fh:
@@ -134,11 +120,12 @@ def save_mappings(selected_container):
     except ValueError:
         # failsafe beta_version
         beta_version = 0
-    l10n_mappings = valid_l10n_mappings()
+    # choose split number
+    l10n_mappings = select_l10n_mappings(beta_version)
     sample_mappings = {k: v for k, v in l10n_mappings.items() if k.startswith("demo")}
     if os.environ.get("TESTRAIL_REPORT") or os.environ.get("MANUAL"):
         # Run all tests if this is a scheduled beta or a manual run
-        save_mappings(distribute_mappings_evenly(l10n_mappings, beta_version))
+        save_mappings(l10n_mappings)
         sys.exit(0)
 
     re_set_all = [
@@ -199,5 +186,5 @@ def save_mappings(selected_container):
                 selected_mappings |= sample_mappings
                 break
 
-    save_mappings(distribute_mappings_evenly(selected_mappings, beta_version))
+    save_mappings(selected_mappings)
     sys.exit(0)

l10n_CM/beta_run_splits/l10n_split_1.json

@@ -0,0 +1,32 @@
+{
+  "demo": [
+    "US",
+    "GB",
+    "DE",
+    "FR",
+    "PL",
+    "ES",
+    "CA"
+  ],
+  "bijoubrigitte": [
+    "DE"
+  ],
+  "calvinklein": [
+    "US"
+  ],
+  "aldoshoes": [
+    "US"
+  ],
+  "peacocks": [
+    "GB"
+  ],
+  "assos": [
+    "GB"
+  ],
+  "fnac": [
+    "FR"
+  ],
+  "canadatire": [
+    "CA"
+  ]
+}

l10n_CM/beta_run_splits/l10n_split_2.json

@@ -0,0 +1 @@
+{"amazon": ["US", "CA", "DE", "FR"], "walmart": ["US", "CA"], "mediamarkt": ["DE"], "zalando": ["PL"], "bestbuy": ["US"], "apple": ["US"], "whittard": ["GB"], "tiffany": ["IT"], "yellowkorner": ["FR"], "burtsbees": ["CA"]}

l10n_CM/beta_run_splits/l10n_split_3.json

@@ -0,0 +1 @@
+{"ebay": ["ES", "IT", "CA", "GB"], "etsy": ["US", "CA"], "justspices": ["DE"], "lowes": ["US"], "newegg": ["US"], "wish": ["US"], "diy": ["GB"], "artsper": ["FR"], "newbalance": ["CA"]}

modules/testrail_integration.py

@@ -4,7 +4,7 @@
 import subprocess
 import sys
 
-from choose_l10n_ci_set import distribute_mappings_evenly, valid_l10n_mappings
+from choose_l10n_ci_set import select_l10n_mappings
 from modules import taskcluster as tc
 from modules import testrail as tr
 from modules.testrail import TestRail
@@ -187,11 +187,9 @@ def reportable(platform_to_test=None):
 
     plan_entries = this_plan.get("entries")
     if os.environ.get("FX_L10N"):
-        report = True
         beta_version = int(minor_num.split("b")[-1])
-        distributed_mappings = distribute_mappings_evenly(
-            valid_l10n_mappings(), beta_version
-        )
+        distributed_mappings = select_l10n_mappings(beta_version)
+        expected_mappings = sum(map(lambda x: len(x), distributed_mappings.values()))
         covered_mappings = 0
         # keeping this logic to still see how many mappings are reported.
         for entry in plan_entries:
@@ -205,13 +203,11 @@ def reportable(platform_to_test=None):
                             and platform in run_platform
                         ):
                             covered_mappings += 1
-                            report = False
         logging.warning(
-            f"Potentially matching run found for {platform}, may be reportable. (Found {covered_mappings} site/region mappings reported.)"
+            f"Potentially matching run found for {platform}, may be reportable. (Found {covered_mappings} site/region mappings reported, expected {expected_mappings}.)"
         )
-        logging.warning(f"Run is reportable: {report}")
-        # Only report when there is a new beta and no other site/region mappings are reported.
-        return report
+        # Only report when there is a new beta without a reported plan or if the selected split is not completely reported.
+        return covered_mappings < expected_mappings
     else:
         covered_suites = 0
         for entry in plan_entries: