style: apply pre-commit black/isort formatting to imported WIP modules

Author: AnkushMalaker

backends/advanced/src/advanced_omi_backend/controllers/background_bucket_controller.py

@@ -31,22 +31,22 @@
 from rq.job import Job
 
 from advanced_omi_backend.constants import BACKGROUND_SPEECH_LABEL, NOISE_LABEL
-from advanced_omi_backend.models.conversation import Conversation
 from advanced_omi_backend.controllers.queue_controller import (
     JOB_RESULT_TTL,
     default_queue,
 )
+from advanced_omi_backend.models.conversation import Conversation
 from advanced_omi_backend.speaker_recognition_client import SpeakerRecognitionClient
 from advanced_omi_backend.users import User
 from advanced_omi_backend.utils.audio_chunk_utils import reconstruct_audio_segment
-from advanced_omi_backend.workers.background_index_jobs import (
-    index_background_corpus_job,
-)
+from advanced_omi_backend.workers.background_benchmark import build_background_benchmark
 from advanced_omi_backend.workers.background_cleanup_jobs import (
     apply_background_cleanup_job,
     build_background_cleanup_report,
 )
-from advanced_omi_backend.workers.background_benchmark import build_background_benchmark
+from advanced_omi_backend.workers.background_index_jobs import (
+    index_background_corpus_job,
+)
 from advanced_omi_backend.workers.background_suppression import zone_for
 
 logger = logging.getLogger(__name__)
@@ -887,9 +887,7 @@ def _novelty_groups(
         ):
             groups.setdefault(row["conversation_id"], []).append(row)
     return [
-        group
-        for group in groups.values()
-        if len(group) >= profile["novelty_min_clips"]
+        group for group in groups.values() if len(group) >= profile["novelty_min_clips"]
     ]
 
 

backends/advanced/src/advanced_omi_backend/controllers/background_suppression_controller.py

@@ -59,7 +59,9 @@ async def get_conversation_suppressions(user: User, conversation_id: str) -> dic
             },
         )
         cluster["segments"].append(doc)
-        cluster["statuses"][doc["status"]] = cluster["statuses"].get(doc["status"], 0) + 1
+        cluster["statuses"][doc["status"]] = (
+            cluster["statuses"].get(doc["status"], 0) + 1
+        )
         cluster["zones"][doc["zone"]] = cluster["zones"].get(doc["zone"], 0) + 1
         cluster["max_background_similarity"] = max(
             cluster["max_background_similarity"],

backends/advanced/src/advanced_omi_backend/controllers/drift_controller.py

@@ -89,7 +89,12 @@ async def drift_fingerprint() -> str:
                             "id": "$$v.version_id",
                             "cent": {
                                 "$cond": [
-                                    {"$ifNull": ["$$v.metadata.cluster_centroids", False]},
+                                    {
+                                        "$ifNull": [
+                                            "$$v.metadata.cluster_centroids",
+                                            False,
+                                        ]
+                                    },
                                     1,
                                     0,
                                 ]
@@ -116,7 +121,11 @@ async def drift_fingerprint() -> str:
     convs = []
     async for doc in Conversation.get_pymongo_collection().aggregate(pipeline):
         active = next(
-            (v for v in doc["versions"] if v["id"] == doc.get("active_transcript_version")),
+            (
+                v
+                for v in doc["versions"]
+                if v["id"] == doc.get("active_transcript_version")
+            ),
             None,
         )
         convs.append(

backends/advanced/src/advanced_omi_backend/controllers/guided_enrollment_controller.py

@@ -206,9 +206,7 @@ async def _candidate_pool(user: User, speaker_name: str, reviewed: set) -> list:
         },
     ):
         segments = _active_segments(doc)
-        speaker_present = any(
-            _effective_label(s) == speaker_name for s in segments
-        )
+        speaker_present = any(_effective_label(s) == speaker_name for s in segments)
         if not speaker_present:
             continue
         audio_duration = doc.get("audio_total_duration") or 0.0

backends/advanced/src/advanced_omi_backend/routers/modules/device_input_routes.py

@@ -73,7 +73,9 @@ def _utc_iso(value: Optional[datetime]) -> Optional[str]:
     return _as_utc(value).isoformat().replace("+00:00", "Z")
 
 
-def _effective_source_status(source: CaptureSource, now: Optional[datetime] = None) -> str:
+def _effective_source_status(
+    source: CaptureSource, now: Optional[datetime] = None
+) -> str:
     if source.status != "online":
         return source.status
     checked_at = _as_utc(now or utcnow())
@@ -242,7 +244,11 @@ async def ingest_activity(
                     start_at=incoming.captured_at,
                     end_at=incoming.ended_at,
                     purpose="timeline_thumbnail",
-                    payload={"item_id": str(item.id), "frame_id": frame_id, "width": 960},
+                    payload={
+                        "item_id": str(item.id),
+                        "frame_id": frame_id,
+                        "width": 960,
+                    },
                 ).insert()
         except DuplicateKeyError:
             duplicates += 1

backends/advanced/src/advanced_omi_backend/workers/background_suppression_jobs.py

@@ -136,9 +136,8 @@ async def backfill_conversation_suppressions(
 
 async def backfill_all_suppressions(user_id: str, limit: int = 500) -> dict:
     """Backfill every conversation present in the corpus index."""
-    conversation_ids = await (
-        _database()["background_corpus_embeddings"]
-        .distinct("conversation_id", {"requested_by": user_id})
+    conversation_ids = await _database()["background_corpus_embeddings"].distinct(
+        "conversation_id", {"requested_by": user_id}
     )
     results = {"conversations": 0, "written": 0, "skipped": 0}
     for conversation_id in conversation_ids[:limit]:

backends/advanced/src/scripts/cleanup_state.py

@@ -243,21 +243,21 @@ def row(label, value, style="white"):
         str(stats.langfuse_prompts),
         "yellow" if stats.langfuse_prompts else "dim",
     )
-    """table.add_section()       #then 
+    """table.add_section()       #then
     row("Redis Jobs", str(stats.redis_jobs), "dim")
     row("Legacy WAV Files", str(stats.legacy_wav), "dim")
     row("Vault Files", str(stats.vault_files), "green" if stats.vault_files else "dim")
     table.add_section()
     row("Users", str(stats.users), "cyan")
 
     return table"""
-    #now
+    # now
     table.add_section()
     row("Redis Jobs", str(stats.redis_jobs), "dim")
     row("Legacy WAV Files", str(stats.legacy_wav), "dim")
     row("Vault Files", str(stats.vault_files), "green" if stats.vault_files else "dim")
     table.add_section()
-    
+
     # ----------------------------------------------------
     # Custom Extension: Disk Storage Metric
     # Added by Omer Said Bayhan (https://github.com/omer-said-bayhan)

backends/advanced/tests/test_background_buckets.py

@@ -3,6 +3,7 @@
 
 from advanced_omi_backend.constants import is_non_enrollable_speaker
 from advanced_omi_backend.controllers.background_bucket_controller import (
+    SURFACE_PROFILES,
     _background_likelihood,
     _cluster_rows,
     _content_signature,
@@ -12,20 +13,18 @@
     _is_known_foreground,
     _max_similarity,
     _novelty_groups,
-    _representatives,
+    _queue_summary,
     _reference_scores,
+    _representatives,
     _review_samples,
-    _queue_summary,
-    SURFACE_PROFILES,
 )
 from advanced_omi_backend.models.conversation import Conversation
 from advanced_omi_backend.routers.modules.annotation_routes import (
     _apply_diarization_label,
 )
-from advanced_omi_backend.workers import speaker_jobs
-from advanced_omi_backend.workers import background_index_jobs
-from advanced_omi_backend.workers.background_cleanup_jobs import _score_rows
+from advanced_omi_backend.workers import background_index_jobs, speaker_jobs
 from advanced_omi_backend.workers.background_benchmark import evaluate_reviews
+from advanced_omi_backend.workers.background_cleanup_jobs import _score_rows
 
 
 def test_gap_windows_sample_audio_without_transcript_segments():
@@ -172,9 +171,21 @@ def test_not_background_examples_suppress_similar_speech_but_not_noise():
 
 def test_confirmed_background_mines_similar_clips_for_batch_harvest():
     rows = [
-        {"clip_key": "a", "conversation_id": "one", "candidate_type": "background_speech"},
-        {"clip_key": "b", "conversation_id": "one", "candidate_type": "background_speech"},
-        {"clip_key": "c", "conversation_id": "two", "candidate_type": "background_speech"},
+        {
+            "clip_key": "a",
+            "conversation_id": "one",
+            "candidate_type": "background_speech",
+        },
+        {
+            "clip_key": "b",
+            "conversation_id": "one",
+            "candidate_type": "background_speech",
+        },
+        {
+            "clip_key": "c",
+            "conversation_id": "two",
+            "candidate_type": "background_speech",
+        },
         {"clip_key": "d", "conversation_id": "one", "candidate_type": "noise"},
     ]
     background = {"a": 0.9, "b": 0.6, "c": 0.4, "d": 0.9}
@@ -188,7 +199,11 @@ def test_confirmed_background_mines_similar_clips_for_batch_harvest():
 
 def test_harvest_requires_background_to_beat_foreground_by_margin():
     rows = [
-        {"clip_key": "a", "conversation_id": "one", "candidate_type": "background_speech"},
+        {
+            "clip_key": "a",
+            "conversation_id": "one",
+            "candidate_type": "background_speech",
+        },
     ]
 
     # familiar foreground voice that also happens to resemble the bucket
@@ -198,13 +213,29 @@ def test_harvest_requires_background_to_beat_foreground_by_margin():
 def test_queue_summary_splits_sign_offs_from_genuine_unknowns():
     clusters = [
         # harvest lane: quick confirm regardless of mean scores
-        {"mined": "harvest", "mean_background_similarity": 0.5, "mean_foreground_similarity": 0.4},
+        {
+            "mined": "harvest",
+            "mean_background_similarity": 0.5,
+            "mean_foreground_similarity": 0.4,
+        },
         # confident zone (>=0.45 and margin >=0.20): quick confirm
-        {"mined": None, "mean_background_similarity": 0.6, "mean_foreground_similarity": 0.3},
+        {
+            "mined": None,
+            "mean_background_similarity": 0.6,
+            "mean_foreground_similarity": 0.3,
+        },
         # unsure band: genuinely uncertain
-        {"mined": None, "mean_background_similarity": 0.45, "mean_foreground_similarity": 0.35},
+        {
+            "mined": None,
+            "mean_background_similarity": 0.45,
+            "mean_foreground_similarity": 0.35,
+        },
         # novelty (low similarity to everything): genuinely uncertain
-        {"mined": "novel", "mean_background_similarity": 0.1, "mean_foreground_similarity": 0.1},
+        {
+            "mined": "novel",
+            "mean_background_similarity": 0.1,
+            "mean_foreground_similarity": 0.1,
+        },
     ]
 
     assert _queue_summary(clusters) == {
@@ -216,31 +247,49 @@ def test_queue_summary_splits_sign_offs_from_genuine_unknowns():
 
 def test_surface_dial_widens_and_narrows_the_harvest_lane():
     rows = [
-        {"clip_key": "a", "conversation_id": "one", "candidate_type": "background_speech"},
+        {
+            "clip_key": "a",
+            "conversation_id": "one",
+            "candidate_type": "background_speech",
+        },
     ]
     # borderline clip: below the default 0.55 floor, above the "more" 0.45 one
     background = {"a": 0.50}
     foreground = {"a": 0.30}
 
     assert _harvest_groups(rows, background, foreground) == []
-    assert _harvest_groups(
-        rows, background, foreground, SURFACE_PROFILES["more"]
-    ) == [[rows[0]]]
+    assert _harvest_groups(rows, background, foreground, SURFACE_PROFILES["more"]) == [
+        [rows[0]]
+    ]
     # "less" tightens past a clip the default would accept
-    assert _harvest_groups(
-        rows, {"a": 0.58}, {"a": 0.30}, SURFACE_PROFILES["less"]
-    ) == []
+    assert (
+        _harvest_groups(rows, {"a": 0.58}, {"a": 0.30}, SURFACE_PROFILES["less"]) == []
+    )
 
 
 def test_novelty_lane_surfaces_clips_unlike_any_labelled_reference():
     rows = [
-        {"clip_key": f"tv{i}", "conversation_id": "tv-conv", "candidate_type": "background_speech"}
+        {
+            "clip_key": f"tv{i}",
+            "conversation_id": "tv-conv",
+            "candidate_type": "background_speech",
+        }
         for i in range(5)
     ] + [
-        {"clip_key": "familiar", "conversation_id": "tv-conv", "candidate_type": "background_speech"},
-        {"clip_key": "lone", "conversation_id": "small-conv", "candidate_type": "background_speech"},
+        {
+            "clip_key": "familiar",
+            "conversation_id": "tv-conv",
+            "candidate_type": "background_speech",
+        },
+        {
+            "clip_key": "lone",
+            "conversation_id": "small-conv",
+            "candidate_type": "background_speech",
+        },
     ]
-    background = {key: 0.1 for key in ("tv0", "tv1", "tv2", "tv3", "tv4", "familiar", "lone")}
+    background = {
+        key: 0.1 for key in ("tv0", "tv1", "tv2", "tv3", "tv4", "familiar", "lone")
+    }
     foreground = {key: 0.1 for key in background} | {"familiar": 0.8}
 
     groups = _novelty_groups(rows, background, foreground, consumed={"tv4"})

backends/advanced/tests/test_background_suppression.py

@@ -47,8 +47,14 @@ def test_same_source_segments_share_a_cluster_signature():
 
 def test_cluster_signatures_are_stable_across_rescoring():
     embeddings = [[1.0, 0.0], [0.98, 0.02]]
-    first = [_record(10.0, "line one", embeddings[0]), _record(20.0, "line two", embeddings[1])]
-    second = [_record(10.0, "line one", embeddings[0]), _record(20.0, "line two", embeddings[1])]
+    first = [
+        _record(10.0, "line one", embeddings[0]),
+        _record(20.0, "line two", embeddings[1]),
+    ]
+    second = [
+        _record(10.0, "line one", embeddings[0]),
+        _record(20.0, "line two", embeddings[1]),
+    ]
     assign_cluster_signatures(first)
     assign_cluster_signatures(second)
     assert first[0]["cluster_signature"] == second[0]["cluster_signature"]

backends/advanced/tests/test_guided_enrollment_pool.py

@@ -34,7 +34,11 @@ def test_placeholder_labels_are_not_attribution():
 
 def _clip(sim, manually_labeled):
     return {
-        "scores": {"sim_centroid": sim, "max_clip_sim": 0.4, "best_other": {"score": 0.1}},
+        "scores": {
+            "sim_centroid": sim,
+            "max_clip_sim": 0.4,
+            "best_other": {"score": 0.1},
+        },
         "duration": 6.0,
         "current_label": "daksh",
         "manually_labeled": manually_labeled,