Fix bug suggestions with PostgreSQL (#7988)

* Use trigram similarity instead of FTS

* Skip escaping special characters

* Order results by match ranking on Postgres
---------

Co-authored-by: Bastien Abadie <bastien@nextcairn.com>
Co-authored-by: Sebastian Hengst <aryx.github@gmx-topmail.de>

tests/log_parser/test_tasks.py

@@ -1,10 +1,11 @@
 import pytest
+from unittest.mock import patch
 
 from tests.test_utils import add_log_response
 from treeherder.etl.jobs import store_job_data
 from treeherder.etl.push import store_push_data
-from treeherder.model.error_summary import get_error_summary
-from treeherder.model.models import Job, TextLogError
+from treeherder.model.error_summary import get_error_summary, bug_suggestions_line
+from treeherder.model.models import Job, TextLogError, Bugscache
 
 from ..sampledata import SampleData
 
@@ -62,3 +63,86 @@ def test_create_error_summary(
     )
     for failure_line in bug_suggestions:
         assert set(failure_line.keys()) == expected_keys
+
+
+@pytest.mark.django_db
+@patch(
+    "treeherder.model.error_summary.get_error_search_term_and_path",
+    return_value={
+        "search_term": ["browser_dbg-pretty-print-inline-scripts.js"],
+        "path_end": "devtools/client/debugger/test/mochitest/browser_dbg-pretty-print-inline-scripts.js",
+    },
+)
+def test_bug_suggestion_line(
+    search_mock, failure_classifications, jobs_with_local_log, sample_push, test_repository
+):
+    """
+    A test to verify similarity of search term (often test name) derived from
+    the failure line and bug summary gets taken into account. If it is equal
+    for every bug, the expected result won't be returned by the query because
+    of its higher bug ID.
+    """
+    store_push_data(test_repository, sample_push)
+    for job in jobs_with_local_log:
+        job["job"]["result"] = "testfailed"
+        job["revision"] = sample_push[0]["revision"]
+    store_job_data(test_repository, jobs_with_local_log)
+
+    job = Job.objects.get(id=1)
+
+    Bugscache.objects.create(
+        id=1775819,
+        status="2",
+        keywords="intermittent-failure,regression,test-verify-fail",
+        whiteboard="[retriggered][stockwell unknown]",
+        summary=(
+            "Intermittent devtools/client/debugger/test/mochitest/browser_dbg-pretty-print-inline-scripts.js "
+            "| single tracking bug"
+        ),
+        modified="2010-01-01 00:00:00",
+    )
+
+    # Create 50 other results with an inferior ID.
+    # The bug suggestions SQL query fetches up to 50 rows, ordered by match rank then ID.
+    # In case results are returned with a wrong rank (e.g. 0 for each result), above related suggestion will be lost.
+    Bugscache.objects.bulk_create(
+        [
+            Bugscache(
+                id=100 + i,
+                status="2",
+                keywords="intermittent-failure,intermittent-testcase",
+                summary=(
+                    f"Intermittent devtools/client/debugger/test/mochitest/browser_unrelated-{i}.js "
+                    "| single tracking bug"
+                ),
+                modified="2010-01-01 00:00:00",
+            )
+            for i in range(50)
+        ]
+    )
+
+    error = job.text_log_error.first()
+    summary, line_cache = bug_suggestions_line(
+        error,
+        project=job.repository,
+        logdate=job.submit_time,
+        term_cache={},
+        line_cache={str(job.submit_time.date()): {}},
+        revision=job.push.revision,
+    )
+    assert summary["bugs"]["open_recent"] == [
+        {
+            "crash_signature": "",
+            "dupe_of": None,
+            "id": 1775819,
+            "keywords": "intermittent-failure,regression,test-verify-fail",
+            "resolution": "",
+            "status": "2",
+            "whiteboard": "[retriggered][stockwell unknown]",
+            "summary": (
+                "Intermittent "
+                "devtools/client/debugger/test/mochitest/browser_dbg-pretty-print-inline-scripts.js "
+                "| single tracking bug"
+            ),
+        }
+    ]

tests/sample_data/bug_list.json

@@ -135,7 +135,7 @@
     {
       "status": "NEW",
       "id": 1054669,
-      "summary": "Intermittent test_switch_frame.py TestSwitchFrame.test_should_be_able_to_carry_on_working_if_the_frame_is_deleted_from_under_us | TimeoutException: TimeoutException: Connection timed out",
+      "summary": "Intermittent test_switch_frame.py TestSwitchFrame.test_should_be_able_to_carry_on_working_if_the_frame_is_deleted_from_under_us | TimeoutException",
       "dupe_of": null,
       "duplicates": [],
       "cf_crash_signature": "",

treeherder/model/migrations/0031_trigram_extension.py

@@ -0,0 +1,13 @@
+# Generated by Django 4.1.13 on 2024-03-25 16:15
+
+from django.db import migrations
+from django.contrib.postgres.operations import TrigramExtension
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("model", "0030_group_durations"),
+    ]
+
+    operations = [TrigramExtension()]

treeherder/model/models.py

@@ -12,12 +12,12 @@ warnings.filterwarnings("ignore", category=DeprecationWarning, module="newrelic"
 import newrelic.agent
 from django.conf import settings
 from django.contrib.auth.models import User
-from django.contrib.postgres.search import SearchQuery, SearchRank, SearchVector
 from django.core.cache import cache
 from django.core.exceptions import ObjectDoesNotExist
 from django.core.validators import MinLengthValidator
 from django.db import models, transaction
 from django.db.models import Count, Max, Min, Q, Subquery
+from django.contrib.postgres.search import TrigramSimilarity
 from django.db.utils import ProgrammingError
 from django.forms import model_to_dict
 from django.utils import timezone
@@ -248,11 +248,11 @@ class Bugscache(models.Model):
     def search(cls, search_term):
         max_size = 50
 
-        # Do not wrap a string in quotes to search as a phrase;
-        # see https://bugzilla.mozilla.org/show_bug.cgi?id=1704311
-        search_term_fulltext = cls.sanitized_search_term(search_term)
-
         if settings.DATABASES["default"]["ENGINE"] == "django.db.backends.mysql":
+            # Do not wrap a string in quotes to search as a phrase;
+            # see https://bugzilla.mozilla.org/show_bug.cgi?id=1704311
+            search_term_fulltext = cls.sanitized_search_term(search_term)
+
             # Substitute escape and wildcard characters, so the search term is used
             # literally in the LIKE statement.
             search_term_like = (
@@ -275,12 +275,16 @@ class Bugscache(models.Model):
                 [search_term_fulltext, search_term_like, max_size],
             )
         else:
-            # On PostgreSQL we can use the full text search features
-            vector = SearchVector("summary")
-            query = SearchQuery(search_term_fulltext)
-            recent_qs = Bugscache.objects.annotate(rank=SearchRank(vector, query)).order_by(
-                "-rank", "id"
-            )[0:max_size]
+            # On PostgreSQL we can use the ORM directly, but NOT the full text search
+            # as the ranking algorithm expects english words, not paths
+            # So we use standard pattern matching AND trigram similarity to compare suite of characters
+            # instead of words
+            # Django already escapes special characters, so we do not need to handle that here
+            recent_qs = (
+                Bugscache.objects.filter(summary__icontains=search_term)
+                .annotate(similarity=TrigramSimilarity("summary", search_term))
+                .order_by("-similarity")[0:max_size]
+            )
 
         exclude_fields = ["modified", "processed_update"]
         try: