text stringlengths 185 73.3k | repo stringlengths 7 100 | path stringlengths 4 146 | language stringclasses 7
values | hash stringlengths 16 16 | score float64 7 8.5 | stars int64 0 237k |
|---|---|---|---|---|---|---|
"""
Tests for the Phase 2.5 MCAP/Foxglove glue.
``export_session_to_mcap`` and ``open_in_foxglove`` are thin wrappers around
``polyumi_ingest``'s real exporter and the local ``foxglove-studio`` binary; those
are exercised elsewhere (ingest's own test suite, and manual smoke testing), so here
we monkeypatch them and te... | cwoodhayes/PolyUMI | catalog/test/test_mcap_tools.py | .py | 36221a934ab31b05 | 7.95 | 7 |
"""Tests for the scene detail pane's recording/preprocessing commit provenance."""
from __future__ import annotations
import json
import pathlib
import zarr
from polyumi_catalog import provenance
def _make_session(scene_dir: pathlib.Path, name: str, polyumi_version: str | None) -> pathlib.Path:
"""
Write a... | cwoodhayes/PolyUMI | catalog/test/test_provenance.py | .py | 08a999c728e013f5 | 7.95 | 7 |
"""Tests for the per-session pzarr stream summary (Phase 4), wrapping ingest's inspect_pzarr."""
from __future__ import annotations
import pathlib
import numpy as np
import zarr
from polyumi_catalog import pzarr_inspect
def _make_pzarr_with_finger_stream(scene_dir: pathlib.Path, session_dirname: str, *, n_frames: ... | cwoodhayes/PolyUMI | catalog/test/test_pzarr_inspect.py | .py | 511ac04f7b094e2a | 7.95 | 7 |
"""Tests for on-demand gopro.mp4 thumbnail decoding (Phase 4)."""
from __future__ import annotations
import pathlib
import cv2
import numpy as np
import pytest
from polyumi_catalog import thumbnails
_N = 40
_H, _W = 48, 64
def _write_test_mp4(path: pathlib.Path, n_frames: int = _N) -> int:
"""Write an n_frame... | cwoodhayes/PolyUMI | catalog/test/test_thumbnails.py | .py | fcfacb7f92fc5ed2 | 7.95 | 7 |
"""
A sine-wave backend, so the ROS client can be brought up without a GPU or a checkpoint.
Two channels oscillate, at the same frequency but 90 degrees apart: X on a sine, gripper width on a
cosine. The phase offset is deliberate -- the gripper's extremes land on X's zero crossings, so a
routing bug that feeds X into... | cwoodhayes/PolyUMI | inference_server/polyumi_inference/backends/sine.py | .py | 1f0e2c0628e8d592 | 7.45 | 7 |
"""
Client half of the inference protocol.
:class:`PolicyClient` is what the ROS-side ``policy_client_node`` holds. It owns the three
endpoints, the URL arithmetic between them, and the persistent connection; it does not own logging
or retry policy, which belong to the caller (the node logs and carries on; a script mi... | cwoodhayes/PolyUMI | inference_server/polyumi_inference/client.py | .py | 83f5773ccf305b35 | 7.45 | 7 |
"""
Exceptions raised across the inference protocol.
Two failure modes, deliberately distinct because they have different fixes. A
:class:`WireFormatError` means the *frame* is wrong -- malformed, truncated, or missing a channel
the policy needs -- and is always the requester's fault, so every server turns it into a 4... | cwoodhayes/PolyUMI | inference_server/polyumi_inference/errors.py | .py | c33f1df12a6f14b8 | 7.45 | 7 |
"""
Server half of the inference protocol: the app every PolyUMI inference server is.
:func:`create_app` owns the HTTP surface -- the three routes, decoding the frame, enforcing the
contract, turning a bad frame into a 422, truncating the chunk, and the request timing. A server is
then only a :class:`PolicyBackend`: a... | cwoodhayes/PolyUMI | inference_server/polyumi_inference/server.py | .py | 01024ad4da12a4ab | 7.45 | 7 |
"""
Drive a real :class:`~polyumi_inference.client.PolicyClient` against a real app, in process.
This is why :class:`~polyumi_inference.client.Transport` is a seam. With it, a test can send the
exact bytes the ROS node sends and have them decoded by the exact code the server runs, with no
socket, no port, and no secon... | cwoodhayes/PolyUMI | inference_server/polyumi_inference/testing.py | .py | f3b27fe7ff6ce4ec | 7.95 | 7 |
"""
The two types the client and the server both hold.
An :class:`Observation` goes one way and an :class:`ActionChunk` comes back, and each is built by
one end and consumed by the other. They are defined once, here, so the two ends cannot disagree
about them -- which is the whole reason this library exists.
Plain da... | cwoodhayes/PolyUMI | inference_server/polyumi_inference/types.py | .py | 1f633addc38e2cb7 | 7.45 | 7 |
"""
Binary framing for the ``/predict_cartesian/`` observation request.
The frame is a length-prefixed JSON header followed by the raw bytes of each channel::
[4B big-endian uint32: header length N][N bytes UTF-8 JSON header][channel blobs]
and the header names every blob's dtype, shape and position::
{"ver... | cwoodhayes/PolyUMI | inference_server/polyumi_inference/wire.py | .py | 5ba710efe1d28da8 | 7.45 | 7 |
"""
Tests for the dummy server's oscillator and its HOME_POSE validation.
The dummy is the only thing exercising the ROS-side action path without a GPU or a checkpoint, so
its output shape matters: if the gripper channel is constant, a broken gripper route looks identical
to a working one.
What it *refuses* is tested... | cwoodhayes/PolyUMI | inference_server/test/test_dummy_server.py | .py | 8dcaf6bee998ed7b | 7.95 | 7 |
"""
Guard the library's Python 3.9 floor.
The diffusion-policy container's conda env is ``python=3.9`` (numpy 1.24) and imports this library;
the ROS node imports it under 3.12. There is no 3.9 interpreter on a development laptop, so the
floor cannot be checked by running the tests -- and the realistic way to break it... | cwoodhayes/PolyUMI | inference_server/test/test_python39_floor.py | .py | 401dfb6ac54a045d | 7.95 | 7 |
r"""
Populate an ORB-SLAM3 settings YAML from OpenImuCameraCalibrator output.
Reads the FISHEYE camera calibration JSON and the cam-IMU calibration result
from a calibration dataset directory, then writes the derived values into the
target ORB-SLAM3 settings YAML in-place.
Supported camera model: FISHEYE (OpenImuCame... | cwoodhayes/PolyUMI | ingest/integration/populate_slam_yaml.py | .py | 82115a8a70efc4e8 | 7.45 | 7 |
"""
Visualize SLAM-to-optitrack alignment for a scene after preprocessing step 3.
Shows three trajectories in the optitrack frame:
1. OptiTrack ground truth (GoPro position, optitrack frame)
2. SLAM trajectory after applying the slam->optitrack transform (aligned)
3. SLAM trajectory before the transform (raw SLA... | cwoodhayes/PolyUMI | ingest/integration/visualize_slam_alignment.py | .py | 5ce249f8e5f89613 | 7.45 | 7 |
"""
Visualize time-sync alignment for a scene after preprocessing step 1 (pp).
Usage:
uv run python ingest/integration/visualize_timesync.py recordings/scene_YYYY-MM-DD_...
"""
import argparse
import logging
import os
import pathlib
import signal
import sys
import matplotlib.pyplot as plt
import matplotlib.ticke... | cwoodhayes/PolyUMI | ingest/integration/visualize_timesync.py | .py | c65cf38f1a896387 | 7.45 | 7 |
"""Resolve BLE devices through Home Assistant."""
from dataclasses import dataclass
import bleak
from bleak import BleakClient
from bleak.backends.device import BLEDevice
from homeassistant.components import bluetooth
from homeassistant.core import HomeAssistant
@dataclass(frozen=True, slots=True)
class BLEDeviceRe... | teh-hippo/ha-govee-led-ble | custom_components/ha_govee_led_ble/ble_device_resolver.py | .py | b0f299a22a777160 | 7.54 | 11 |
"""Constants for HA Govee LED BLE."""
from collections.abc import Mapping
from dataclasses import dataclass
from typing import Any
DOMAIN = "ha_govee_led_ble"
CONF_MODEL = "model"
CONF_EFFECT_CATEGORIES = "effect_categories"
CONF_EFFECT_FAMILIES = "effect_families"
CONF_PREFIX_EFFECT_NAMES = "prefix_effect_names"
CON... | teh-hippo/ha-govee-led-ble | custom_components/ha_govee_led_ble/const.py | .py | eb7f53541f1cc059 | 7.54 | 11 |
"""Shared typed base for the coordinator and its write mixins."""
from collections.abc import Awaitable, Callable, Sequence
from typing import TYPE_CHECKING, Any
from homeassistant.helpers.update_coordinator import DataUpdateCoordinator
from .const import ModelProfile
from .control_arbiter import BLEControlArbiter, ... | teh-hippo/ha-govee-led-ble | custom_components/ha_govee_led_ble/coordinator_base.py | .py | 82e86b3c3de50fa4 | 7.54 | 11 |
"""Versioned document storage port and Home Assistant adapter."""
from __future__ import annotations
from collections.abc import Awaitable, Callable
from typing import Any, Protocol
from homeassistant.core import HomeAssistant
from homeassistant.helpers.storage import Store
type EffectDocument = dict[str, Any]
type... | teh-hippo/ha-govee-led-ble | custom_components/ha_govee_led_ble/effect_store.py | .py | 33766cd697d8e687 | 7.54 | 11 |
"""Shared base entity for the Govee BLE integration."""
from homeassistant.helpers.update_coordinator import CoordinatorEntity
from .coordinator import GoveeBLECoordinator
class GoveeBLEEntity(CoordinatorEntity[GoveeBLECoordinator]):
"""Base entity whose availability follows BLE presence and the coordinator ref... | teh-hippo/ha-govee-led-ble | custom_components/ha_govee_led_ble/entity.py | .py | 871c20dbebba2bb2 | 7.54 | 11 |
"""
Lazarus Forge - Shared Audit Library (audit_lib.py)
Extracted from Automation/AUDIT_HARNESS.py, 2026-07-21, so that
Automation/integrity_check.py (repo-wide checks) and AUDIT_HARNESS.py
(single-session audits) both import one implementation of routing
parsing, cross-reference extraction, and finding classification... | ksarith/LazarusForge | Automation/audit_lib.py | .py | b1ab7f289d938b74 | 7.63 | 17 |
"""BB content research strategy — wraps fetch_ai_creators + parse_extract."""
from __future__ import annotations
import json
import logging
import re
import time
from datetime import datetime, timezone
from pathlib import Path
from typing import Any
from genlab_core.strategies import ContentResearchStrategy
BB_ROOT... | AnarchistSid/genlab-platform | BlackboxBrief/bb_strategies/content_research.py | .py | fc3a7c0132ea3e6b | 7.45 | 7 |
"""BB platform adaptation strategy.
Migrated to BasePlatformAdaptationStrategy (Sprint 69). Inherits full
8-rule platform enforcement for all 6 platforms.
Adds (OPS #2 + #6, 2026-06-30) the source-tool affiliate disclosure
hook — after the standard per-platform adaptations run, look at the
SOURCE video's metadata and... | AnarchistSid/genlab-platform | BlackboxBrief/bb_strategies/platform_adaptation.py | .py | d0efcd60b705efc5 | 7.45 | 7 |
"""BB scoring strategy — wraps dedupe_rank_items.py."""
from __future__ import annotations
import json
import logging
import time
from datetime import datetime, timezone
from pathlib import Path
from typing import Any
from genlab_core.strategies import ScoringStrategy
BB_ROOT = Path(__file__).resolve().parent.paren... | AnarchistSid/genlab-platform | BlackboxBrief/bb_strategies/scoring.py | .py | b730218492ac1e31 | 7.45 | 7 |
"""BB writing strategy — LLM content generation for AI creator stories.
Migrated to BaseWritingStrategy (Sprint 69). Inherits the shared LLM +
template-fallback writing pipeline with AI-specific config from
config/writing.yaml and config/templates.yaml.
Previous version (115 lines) called write_video_content() direct... | AnarchistSid/genlab-platform | BlackboxBrief/bb_strategies/writing.py | .py | f40393e156409dcc | 7.45 | 7 |
"""YouTube connector for creator-video discovery."""
from __future__ import annotations
import os
import re
from datetime import datetime, timedelta, timezone
from typing import Any, Dict, Iterable, List, Optional
def _build_service():
"""Build YouTube API service via OAuth refresh token credentials."""
cli... | AnarchistSid/genlab-platform | BlackboxBrief/execution/sources/youtube_connector.py | .py | 500c2e8890555d4a | 7.45 | 7 |
"""Centralized YAML config loader with caching.
Provides a single function to load any config file from config/ directory.
Caches loaded configs in memory to avoid repeated disk reads within a run.
Shared-config routing
---------------------
A handful of yamls are owned by ``genlab-core`` and read by every niche
(Sh... | AnarchistSid/genlab-platform | BlackboxBrief/execution/utils/config_loader.py | .py | b47cb63d13745365 | 7.45 | 7 |
"""Typed boundary contracts for backlog and artifact payloads."""
from __future__ import annotations
from dataclasses import dataclass
from typing import Any, Dict
class ContractValidationError(ValueError):
"""Raised when a boundary payload does not satisfy contract."""
@dataclass(frozen=True)
class Blueprint... | AnarchistSid/genlab-platform | BlackboxBrief/execution/utils/contracts.py | .py | 382cc4a1bf3f2bd2 | 7.45 | 7 |
"""Structured error event helpers for operational pipeline logging."""
from __future__ import annotations
import json
from dataclasses import dataclass, field
from datetime import datetime, timezone
from typing import Any, Dict, Mapping, Optional
@dataclass(frozen=True)
class ErrorEvent:
"""Normalized machine-r... | AnarchistSid/genlab-platform | BlackboxBrief/execution/utils/error_events.py | .py | 4234a3fc680b399c | 7.45 | 7 |
"""Playwright-based fetcher for JS-rendered pages.
Some sources (OpenAI blog, Anthropic news) serve JS-rendered SPAs that
feedparser/requests can't parse. This module uses Playwright to render
the page and extract article content.
Requires: pip install playwright && python -m playwright install chromium
Falls back g... | AnarchistSid/genlab-platform | BlackboxBrief/execution/utils/playwright_fetcher.py | .py | a1e09820d0f9bea5 | 7.45 | 7 |
"""Daily cleanup of old pipeline runs using the DiskQuotaManager.
Keeps the 5 most recent runs protected, evicts lowest-scoring runs first.
Designed to run via launchd at 00:30 UTC (06:00 IST) daily — before the
daily pipeline fires so it never competes for disk space mid-render.
Retention windows (optimised for 1-po... | AnarchistSid/genlab-platform | BlackboxBrief/scripts/cleanup_runs.py | .py | ca07b246e859dc48 | 7.45 | 7 |
"""Tests for the updated score_authority() function.
Verifies that source_priority (from sources.yaml) takes precedence
over domain-based lookup, with proper fallback behavior.
"""
from execution.dedupe_rank_items import score_authority
AUTHORITY_MAP = {
"reddit.com": 0.85,
"i.redd.it": 0.85,
"youtube.co... | AnarchistSid/genlab-platform | BlackboxBrief/tests/_legacy/test_authority_scoring.py | .py | 25d88774f3698f14 | 7.95 | 7 |
"""Shared pytest fixtures for the content intelligence pipeline."""
import json
import os
from unittest.mock import MagicMock
import pytest
# Disable Postgres so tests use SharePoint mocks (avoid schema gaps).
# Must set before any genlab_core import triggers pydantic-settings.
os.environ["GENLAB_USE_POSTGRES"] = "f... | AnarchistSid/genlab-platform | BlackboxBrief/tests/conftest.py | .py | 39f41d432dc9e704 | 7.95 | 7 |
#!/usr/bin/env python3
"""Integration tests for BacklogClient (Microsoft Lists).
These tests require a live Microsoft Graph connection AND an explicit opt-in
via RUN_INTEGRATION_TESTS=1 environment variable. This prevents accidental
execution against production data during normal test runs.
SAFETY: These tests hit RE... | AnarchistSid/genlab-platform | BlackboxBrief/tests/test_backlog_integration.py | .py | de204a24ebdc0c91 | 7.95 | 7 |
"""Tests for file-based cache."""
import json
import time
import pytest
from genlab_core.cache.disk_cache import Cache
@pytest.fixture
def cache(tmp_path):
"""Create a cache in a temp directory."""
return Cache(str(tmp_path / "cache"))
class TestCache:
def test_set_and_get(self, cache):
cache.... | AnarchistSid/genlab-platform | BlackboxBrief/tests/test_cache.py | .py | c7b7a1fd327dd0e1 | 7.95 | 7 |
"""Tests for the publish lock mechanism in cleanup_runs.py."""
from __future__ import annotations
import sys
from pathlib import Path
# Ensure scripts/ is importable
PROJECT_ROOT = Path(__file__).resolve().parent.parent
sys.path.insert(0, str(PROJECT_ROOT / "scripts"))
from cleanup_runs import (
acquire_publish... | AnarchistSid/genlab-platform | BlackboxBrief/tests/test_cleanup_publish_lock.py | .py | 1a262bd83e7f9f87 | 7.95 | 7 |
"""Regression guards for ``execution.utils.config_loader`` after the
symlink-to-shared-config refactor.
History
-------
``BlackboxBrief/config/`` used to contain seven symlinks pointing at
``genlab-core/config/*.yaml``. They were the last "BB-as-special-case"
architectural debt: every reader had to know about the link... | AnarchistSid/genlab-platform | BlackboxBrief/tests/test_config_loader.py | .py | a668053406f127c3 | 7.95 | 7 |
"""Tests for execution/utils/ffmpeg_utils.py constants and helpers.
P1.5: FINAL_VIDEO_PARAMS must include -r 30 for consistent framerate.
P1.6: No hardcoded 192k audio bitrate anywhere — canonical value is 256k.
P1.14: Clip pre-processing timeout must scale with duration.
"""
# ── P1.5: FINAL_VIDEO_PARAMS must con... | AnarchistSid/genlab-platform | BlackboxBrief/tests/test_ffmpeg_utils.py | .py | 98ae2939ee0f61c0 | 7.95 | 7 |
"""Tests for backlog formula → OData $filter translator.
Validates all 7 formula patterns used across the codebase.
"""
from genlab_core.http.graph_proxy import formula_to_odata as _formula_to_odata
class TestSimpleEquality:
"""Pattern 1: {field}='value' → fields/field eq 'value'"""
def test_simple_string(... | AnarchistSid/genlab-platform | BlackboxBrief/tests/test_formula_translator.py | .py | d6e958e1530c58ad | 7.95 | 7 |
"""Dedup capability — find & merge duplicate entities through the agent (COG-122).
Reuses the EXISTING entity-resolution engine end-to-end (no reimplementation,
no new matching code):
* ``plan`` proposes ONE :class:`PlanStep` (``capability="dedup"``,
``action="run_dedup"``) describing a second-pass entity-resolutio... | infona-ai/infona-oss | infona_client/agent/capabilities/dedup_cap.py | .py | 73304ac600ba5e74 | 7.42 | 6 |
"""Brief scope / attribute-match clarify steps for the enrich capability.
When a described subset, multi-value scope, or value-filter cannot be
resolved — or a weak schema-attr mapping needs approval — emit a short
``action="clarify"`` step instead of a silent whole-type enrich or an
empty paid job.
Invariants other ... | infona-ai/infona-oss | infona_client/agent/capabilities/enrich_clarify.py | .py | 9eb2ed77dcbe2e71 | 7.42 | 6 |
"""Shared constants and call-time host lookup for the enrich capability.
Owns process-wide background-task strong-refs (``_bg_tasks`` / ``_spawn``)
and the small scope-value splitter. Implementation of plan/execute lives in
sibling ``enrich_*.py`` modules.
Invariants other agents must not break:
- Look up monkeypatch... | infona-ai/infona-oss | infona_client/agent/capabilities/enrich_common.py | .py | d089912bb53111e4 | 7.42 | 6 |
"""Tier, source-clause, and plan-time cost estimate for enrich.
Owns ``_coerce_tier``, the free-registry coverage probe, the human
"via …" clause, and the honest paid-call estimate (COG-123).
Invariants other agents must not break:
- Per-entity paid cost / has_paid come from adapter-declared metadata
(``_resolve_ch... | infona-ai/infona-oss | infona_client/agent/capabilities/enrich_cost.py | .py | 8a5c3a052e4b9fbb | 7.42 | 6 |
"""Execute a planned enrichment as a background EnrichJob.
Owns count / subset-URI / multi-value-scope resolution helpers and
:meth:`EnrichCapability.execute`. Builds the same :class:`EnrichJob` the
``/enrich/jobs`` route builds and hands it to the shared executor.
Invariants other agents must not break:
- Look up ``... | infona-ai/infona-oss | infona_client/agent/capabilities/enrich_execute.py | .py | 6ce73a3699116c6f | 7.42 | 6 |
"""LLM + deterministic extraction of an EnrichRequest from NL.
Owns the schema-grounded extract prompt, the OpenRouter call, JSON
coercion, and the no-key regex fallback parser.
Invariants other agents must not break:
- Look up ``openrouter_chat`` and ``logger`` on the public ``enrich_cap``
module via :func:`_host`... | infona-ai/infona-oss | infona_client/agent/capabilities/enrich_extract.py | .py | 06587beb45399d6a | 7.42 | 6 |
"""Refresh / overwrite / composite-scope detectors for the enrich capability.
Owns conflict-policy selection (stage / verify / overwrite) and the
conservative verb detectors that flip a plan onto the refresh or
replace rail.
Invariants other agents must not break:
- A bare "refresh / re-verify" stays ``verify`` (ONTA... | infona-ai/infona-oss | infona_client/agent/capabilities/enrich_intent.py | .py | 3d60617300e86b7b | 7.42 | 6 |
"""Schema-ground a parsed enrich request (attributes / scope / subset / tier).
Owns sanitizing an extracted request against the type's real schema, plus
the attribute-name normalizer and the confidence-gated soft matcher
(exact / high-similarity auto-accept; weaker unique hits need approval).
Invariants other agents ... | infona-ai/infona-oss | infona_client/agent/capabilities/enrich_validate.py | .py | 2d65165ce999ef94 | 7.42 | 6 |
"""Cleanup capability — propose + apply normalization rules through the agent.
Reuses the existing normalization engine end-to-end (no reimplementation):
* ``plan`` → :func:`infona_client.normalization.inference.suggest_rules_for_predicates`
(the TARGETED variant: only the predicate(s) the instruction names, never ... | infona-ai/infona-oss | infona_client/agent/capabilities/normalize_cap.py | .py | d47bdad435767999 | 7.42 | 6 |
"""Read-only Q&A capability — wraps the existing NL→SPARQL ask pipeline.
This is the only capability that needs no plan/confirm round-trip: a question
does not mutate the graph, so the agent answers immediately. The planner
special-cases the ``question`` intent and calls :meth:`QueryCapability.answer`
directly. We sti... | infona-ai/infona-oss | infona_client/agent/capabilities/query.py | .py | f163c43b6f941f27 | 7.42 | 6 |
"""Subscribe capability — set up a recurring standing alert / weekly refresh
through the agent (ONTA-235).
This is the conversational front door to the ``notify`` schedule: the persona
says "set up a standing weekly alert that notifies my orchestrator whenever a
model I route to changes price or gets a deprecation dat... | infona-ai/infona-oss | infona_client/agent/capabilities/subscribe_cap.py | .py | 0d124cf60f9813db | 7.42 | 6 |
"""Web-research capability — answer a question FROM THE WEB, cite it, don't store it.
This is the READ-ONLY web-research capability. Premium web-discovery ingest
INGEST creates new graph entities from a query; research ANSWERS a question by
reading the web and returns a cited answer plus a downloadable table
(CSV/JSON... | infona-ai/infona-oss | infona_client/agent/capabilities/web_research_cap.py | .py | 87920ef9e71235f1 | 7.42 | 6 |
"""Intent classifier (one bounded LLM call) for the agent planner.
Looks up ``openrouter_chat`` / ``get_capabilities`` on the
:mod:`infona_client.agent.planner` facade at call time so existing
monkeypatches keep working.
"""
from __future__ import annotations
import json
import structlog
from infona_client.agent.co... | infona-ai/infona-oss | infona_client/agent/planner_classify.py | .py | 8a89a0380279ce5c | 7.42 | 6 |
"""Confirm / execute path for the agent planner.
Looks up ``get_capability`` / ``order_steps`` / ``register_capability``
on the :mod:`infona_client.agent.planner` facade at call time so existing
monkeypatches keep working.
"""
from __future__ import annotations
import os
from datetime import datetime, timedelta, time... | infona-ai/infona-oss | infona_client/agent/planner_execute.py | .py | 335ffec5077dc4f2 | 7.42 | 6 |
"""Classify → dispatch: handle / _respond.
Looks up ``openrouter_chat`` / ``get_capability`` / ``order_steps`` /
``_classify`` on the :mod:`infona_client.agent.planner` facade at call
time so existing monkeypatches keep working. The hosted-only web-ingest
discover path stays here (web_ingest is not in OSS).
"""
from _... | infona-ai/infona-oss | infona_client/agent/planner_handle.py | .py | ffa9a9e3f297f92b | 7.42 | 6 |
"""Session-transcript helpers for the agent planner.
Implementation sibling of :mod:`infona_client.agent.planner`.
"""
from __future__ import annotations
import re
from infona_client.agent.conversation_store import Turn
# How many recent turns of a (possibly long, history-backed) transcript to feed
# the classifier... | infona-ai/infona-oss | infona_client/agent/planner_history.py | .py | 00c4c9a80f7deced | 7.42 | 6 |
"""Capability protocol + module-level registry for the unified Ask-AI agent.
The agent has exactly ONE conversational surface (``POST /graphs/{tenant}/agent``).
Everything a user can ask the agent to do is a *capability* registered here —
question answering, normalization, enrichment, and (later) dedup / ontology
edit... | infona-ai/infona-oss | infona_client/agent/registry.py | .py | 186cea2e5567e041 | 7.42 | 6 |
from fastapi import Request
from infona_client.enrichment.cache import get_enrichment_cache
from infona_client.enrichment.executor import EnrichmentExecutor
from infona_client.enrichment.job_store import make_job_store
from infona_client.enrichment.sources.wikidata import WikidataAdapter
from infona_client.graph.clien... | infona-ai/infona-oss | infona_client/api/deps.py | .py | 7b046df74d959a02 | 7.42 | 6 |
"""THE single conversational surface for the unified Ask-AI agent (COG-118).
``POST /graphs/{tenant}/agent`` is the ONLY endpoint. Everything the agent can do
is a capability behind the registry — there is no per-task conversational
endpoint. The legacy ``/ask``, ``/enrich/*`` and ``/normalize/*`` routes stay for
back... | infona-ai/infona-oss | infona_client/api/routes/agent.py | .py | 635b267bc999d2ff | 7.42 | 6 |
import time
import structlog
from fastapi import APIRouter, Depends, HTTPException, Request
from infona_client.analytics import distinct_id_for, emit
from infona_client.api.deps import get_enrichment_job_store, get_neptune_client
from infona_client.api.rate_limit import limiter
from infona_client.auth.api_keys import... | infona-ai/infona-oss | infona_client/api/routes/ask.py | .py | 30901b1d3f7377ab | 7.42 | 6 |
"""POST /graphs/{tenant}/corrections — the A10 user-correction write path (ONTA-281).
The ONE canonical route every client (Explorer webapp, CLI, MCP) uses to fix a
wrong fact. A human corrects an attribute value in the Explorer; this route turns
it into an A10 :class:`~infona_client.pipeline.corrections.UserAssertion... | infona-ai/infona-oss | infona_client/api/routes/corrections.py | .py | 12de2934bdff2ff3 | 7.42 | 6 |
"""HTTP routes for the auto-enrichment feature."""
from __future__ import annotations
import asyncio
import os
import uuid
from datetime import datetime, timezone
from typing import Optional
from fastapi import APIRouter, Depends, HTTPException, Query
from pydantic import BaseModel
from infona_client.api.deps impor... | infona-ai/infona-oss | infona_client/api/routes/enrich.py | .py | a8f045ace0266dc7 | 7.42 | 6 |
"""Entity detail, Explorer search, and ER-rebuild routes.
``er_rebuild`` is a real merge via ``rewrite_subject`` plus one
``refresh_after_write`` — do not fork a second write path.
"""
from __future__ import annotations
import asyncio
from typing import Any
import structlog
from fastapi import Depends, Query
from ... | infona-ai/infona-oss | infona_client/api/routes/explore_entity.py | .py | 7bd1912aa9f08a47 | 7.42 | 6 |
"""ACP auth helpers — detect and advertise Nastech authentication methods."""
from __future__ import annotations
from typing import Any, Optional
TERMINAL_SETUP_AUTH_METHOD_ID = "nastech-setup"
def detect_provider() -> Optional[str]:
"""Resolve the active Nastech runtime provider, or None if unavailable.
... | nastechresearch/nastech-agent | acp_adapter/auth.py | .py | d384e1b33d3b78b1 | 7.5 | 9 |
"""CLI entry point for the nastech-agent ACP adapter.
Loads environment variables from ``~/.nastech/.env``, configures logging
to write to stderr (so stdout is reserved for ACP JSON-RPC transport),
and starts the ACP agent server.
Usage::
python -m acp_adapter.entry
# or
nastech acp
# or
nastech-... | nastechresearch/nastech-agent | acp_adapter/entry.py | .py | 28f680d3925071bf | 7.5 | 9 |
"""ACP permission bridging for Nastech dangerous-command approvals."""
from __future__ import annotations
import asyncio
import logging
from concurrent.futures import TimeoutError as FutureTimeout
from itertools import count
from typing import Callable
from acp.schema import (
AllowedOutcome,
PermissionOptio... | nastechresearch/nastech-agent | acp_adapter/permissions.py | .py | 524aa20df123093d | 7.5 | 9 |
"""Derive ACP session-provenance metadata from the existing compression chain.
This is an additive Nastech extension surfaced under ACP ``_meta.nastech`` so
existing ACP clients ignore it. It carries no new persisted state: everything
is derived on demand from the ``sessions`` table (``parent_session_id`` /
``end_reas... | nastechresearch/nastech-agent | acp_adapter/provenance.py | .py | 59b145a7aecbccbd | 7.5 | 9 |
"""OpenAI-shape bridge shared by Nastech' ACP clients.
An ACP agent (``copilot --acp``, and the ACP CLIs that reach Nastech as
providers) speaks the Agent Client Protocol, which has no OpenAI-style
``tools``/``tool_calls`` channel: a prompt is text, and a response is text plus
the agent's *own* tool notifications. Nas... | nastechresearch/nastech-agent | agent/acp_openai_bridge.py | .py | 5f11081902b9922b | 7.5 | 9 |
"""Ambient session-accounting context for auxiliary LLM calls.
Auxiliary calls (vision, compression, title generation, web_extract,
session_search, ...) funnel through ``agent.auxiliary_client`` which has no
session handle — so their token usage was historically discarded, leaving
dashboard analytics blind to aux mode... | nastechresearch/nastech-agent | agent/aux_accounting.py | .py | 4d1e6c36c76187a7 | 7.5 | 9 |
"""Single owner for backend identity and failure-scoped skip decisions.
Every fallback / dedup / skip / quarantine decision in Nastech ultimately asks
one question: **"is this candidate the same backend as the one that failed,
along the axis that failure invalidated?"** Before this module, that
question was re-implem... | nastechresearch/nastech-agent | agent/backend_identity.py | .py | d56ef72cf0bf57e0 | 7.5 | 9 |
"""System-battery read-out for the CLI/TUI status bar.
Reads the host battery through ``psutil`` (already a Nastech dependency) and
exposes a compact, colour-coded label. Everything degrades to "unavailable"
when there is no battery (desktops, servers, VMs) or when the read fails, so
callers can render the result unc... | nastechresearch/nastech-agent | agent/battery.py | .py | 72a8c442e52987ed | 7.5 | 9 |
"""Provider-agnostic billing/credit recovery links.
Maps a billing-classified failure onto a recovery link + label. *Detection*
is not done here — that is :mod:`agent.error_classifier`
(``FailoverReason.billing``), the single source of truth for "credit wall vs.
rate limit / auth / transport". The resulting :class:`Bi... | nastechresearch/nastech-agent | agent/billing_links.py | .py | 43c614fe854066f3 | 7.5 | 9 |
"""Bounded reads of HTTP error response bodies.
When a provider returns a non-OK status on a *streaming* request, Nastech reads
the response body to build a useful diagnostic error. A bare ``response.read()``
on a streaming httpx response is unbounded in two dangerous ways:
1. A server can declare (or stream) an arbi... | nastechresearch/nastech-agent | agent/bounded_response.py | .py | e1a3123404e3f27b | 7.5 | 9 |
"""Mint a provider API key by running a command (``key_cmd``).
Static API keys are the exception at enterprise gateways: SSO/OIDC brokers,
cloud IAM, and internal auth proxies all issue SHORT-LIVED bearers instead.
A key copied into ``.env`` (``key_env``) is stale within the hour, so every
request after that 401s and ... | nastechresearch/nastech-agent | agent/command_token_source.py | .py | 7bfde21c9de08f29 | 7.5 | 9 |
# custom_components/evconduit/abrp.py
"""ABRP (A Better Route Planner) telemetry client."""
import logging
import time
import aiohttp
from .const import ABRP_API_URL
_LOGGER = logging.getLogger(__name__)
class ABRPClient:
"""Client for sending telemetry to ABRP."""
def __init__(self, session: aiohttp.Cli... | stevelea/evconduit-homeassistant | custom_components/evconduit/abrp.py | .py | dd7c5f365a8a8eb0 | 7.57 | 13 |
from homeassistant import config_entries
from homeassistant.helpers.selector import EntitySelector, EntitySelectorConfig
import voluptuous as vol
import logging
from .const import (
DOMAIN, CONF_API_KEY, CONF_VEHICLE_ID, CONF_UPDATE_INTERVAL,
CONF_ENVIRONMENT, CONF_ABRP_TOKEN, CONF_ODOMETER_ENTITY,
CONF_ELE... | stevelea/evconduit-homeassistant | custom_components/evconduit/config_flow.py | .py | a57f93eef9c9d1ef | 7.57 | 13 |
# custom_components/evconduit/device_tracker.py
import logging
from homeassistant.components.device_tracker import SourceType, TrackerEntity
from homeassistant.helpers.entity import DeviceInfo
from homeassistant.helpers.update_coordinator import CoordinatorEntity
from .const import DOMAIN
from .sensor import _build_d... | stevelea/evconduit-homeassistant | custom_components/evconduit/device_tracker.py | .py | 826fa664ff873444 | 7.57 | 13 |
# custom_components/evconduit/image.py
"""Image platform exposing the vehicle's manufacturer artwork."""
import logging
from homeassistant.components.image import ImageEntity
from homeassistant.core import callback
from homeassistant.helpers.entity import DeviceInfo
from homeassistant.helpers.update_coordinator impo... | stevelea/evconduit-homeassistant | custom_components/evconduit/image.py | .py | 2c8162bce486c551 | 7.57 | 13 |
"""v7.7 Binding Field — composite field from live interconnect + buffer + sense.
Named from live evidence:
- Local spectral/graph coupling can be high while epoch binding lags (UNBOUND).
- Resonant Frame live buffer can hold samples without closed frames (BUFFER_PENDING).
- Until both binding and frames warm, se... | jacksonjp0311-gif/Cortex | cortex/binding_field.py | .py | 9da0cf96cf105c70 | 7.66 | 20 |
"""SPIKE: audio-driven H3 — pin a full audio TRACK across the whole clip.
Appended to ComfyUI-H3-Motion-Context/nodes.py inside the worker container.
Reuses the pack's layout/payload patches and audio encode path: the track is
encoded with the H3 audio VAE and its reference rows are TRANSLATED onto the
clip's own time... | pizzato/stephen_spielbot | docker/comfyui/h3_audio_track/append_to_motion_context.py | .py | 73c17202a243501e | 7.45 | 7 |
"""C2PA / Content Credentials signing for published videos.
Embeds a cryptographically-signed provenance manifest declaring the video as
AI-generated by Stephen Spielbot. Best-effort: when c2patool is not installed
(or signing fails) we log and skip, exactly like the rest of the media pipeline.
Signing must be the LA... | pizzato/stephen_spielbot | pipeline/c2pa.py | .py | 3cbe54b53552ef75 | 7.45 | 7 |
"""Build a timed caption (SRT) track from a finished film's work directory.
The spoken text is known exactly — narration is the script we fed to TTS, an
acted scene's dialogue is the lines the take performed, and a singing scene's
lyrics are the slice of the song it mouths — so we can hand YouTube accurate
subtitles i... | pizzato/stephen_spielbot | pipeline/captions.py | .py | 958fd3b878cdf03c | 7.45 | 7 |
"""Chatterbox Multilingual TTS — the multilingual narration engine (issue #176).
Resemble AI's Chatterbox Multilingual synthesises expressive speech in 23
languages with zero-shot voice cloning from a reference clip, so the existing
voice library carries straight over. Code (``chatterbox-tts``) and weights
(``Resemble... | pizzato/stephen_spielbot | pipeline/chatterbox.py | .py | 11a7b5d129be5d85 | 7.45 | 7 |
"""YouTube cover image generation utilities (shared between app.py and resume_generation.py)."""
from __future__ import annotations
import logging
from pathlib import Path
from pipeline import prompts as _prompts
logger = logging.getLogger("video_gen")
COVER_WIDTH = 1280
COVER_HEIGHT = 720
def cover_dimensions(v... | pizzato/stephen_spielbot | pipeline/cover.py | .py | f2dbb96f6ba9c229 | 7.45 | 7 |
"""Learned durations for post-render film-edit jobs.
The main render learns how long each task kind takes and predicts an ETA
(see ``pipeline/timing.py`` + ``DurableStore.timing_table``). The film-edit
operations — scene re-renders, the parallel scene upscale, narrator/music
regeneration — run as in-process daemon th... | pizzato/stephen_spielbot | pipeline/film_timing.py | .py | 3ebc889ce0ad3a4e | 7.45 | 7 |
"""Background-music regeneration history.
Music is a film-level asset (one ``background_music.wav`` per work dir), so this
mirrors the cover-image history rather than the per-scene image/video history:
there is no scene id. Every regeneration keeps the prior tracks so the user can
listen to each one and pick the best.... | pizzato/stephen_spielbot | pipeline/music_history.py | .py | b69da248f3d2d6f1 | 7.45 | 7 |
"""Where an acted take can be picked up again.
H3 can only continue a take it still holds the motion context for, and that
context is a latent file in the WORKER's ComfyUI output folder — not in the film
folder, and not on any other worker. So each acted render drops a note next to
the scene saying where its continuat... | pizzato/stephen_spielbot | pipeline/scene_context.py | .py | d7a9a26cce79f18b | 7.45 | 7 |
"""Quality gate for performance shots: did they say the line?
H3 is stochastic — a shot can come out with the wrong words, invented tail
speech, or garbled delivery, and until now nothing checked. The gate
transcribes each rendered shot (faster-whisper, CPU, ~2s per clip against a
~6 minute render) and scores the tran... | pizzato/stephen_spielbot | pipeline/shot_gate.py | .py | 3934bb068357180a | 7.45 | 7 |
"""Where the singing actually is inside a generated song.
A music video pins each scene's slice of the track into its H3 take, so the
prompt has to AGREE with that slice. Two ways it used not to: naming lyrics
that are not in the slice, and asking for a moving mouth over an instrumental
intro. Both pull the model off ... | pizzato/stephen_spielbot | pipeline/song_timing.py | .py | 28525a8a0cc12b34 | 7.45 | 7 |
"""Per-style look of burned-in subtitles (open captions).
The burn itself is ffmpeg's ``subtitles`` filter (libass). Left alone it draws
every film the same way — Arial, white, bottom-centre. A style's
``subtitle_style`` dict (font, size, colours, outline, box, position) is turned
into an ASS ``force_style`` override ... | pizzato/stephen_spielbot | pipeline/subtitle_style.py | .py | c817a1479e5da926 | 7.45 | 7 |
"""Zero-shot singing-voice conversion — "Sing this as [voice]".
seed-vc re-voices a SUNG track as any library voice from its ~10 s reference
clip: melody, timing and words are kept, only the timbre changes. It is the
one true voice-clone in the singing pipeline — the music engines can only be
*described* a vocalist, n... | pizzato/stephen_spielbot | pipeline/svc.py | .py | 378a6b35fa626c7c | 7.45 | 7 |
"""Render-ETA prediction from learned per-task durations.
The orchestrator records how long each task kind takes (see
``DurableStore.timing_table``). This module turns that learned table, plus a
job's planned tasks and the configured worker counts, into a wall-clock ETA.
Model (deliberately simple — see CLAUDE.md §2... | pizzato/stephen_spielbot | pipeline/timing.py | .py | 39047f176b15d43d | 7.45 | 7 |
"""TTS-engine registry: selectable narration models (mirrors pipeline/engines.py).
A TTS *engine* is a narration model the user can pick per style as
``tts_engine`` (falls back to the Apache-2.0 default). All engines do
zero-shot voice cloning from a reference WAV, so the voice library carries
across. Each engine name... | pizzato/stephen_spielbot | pipeline/tts_engines.py | .py | 078e8e0bc7dc81d8 | 7.45 | 7 |
#!/usr/bin/env python3
"""HTTP F5-TTS server — the containerized TTS worker (issue #12).
Wraps the F5-TTS CLI in a tiny HTTP service so the TTS worker can run as a
container with no SSH access.
Endpoints:
GET /health -> {"status": "ok"}
POST /tts -> audio/wav bytes
body: {"text": "... | pizzato/stephen_spielbot | pipeline/tts_server.py | .py | adc9911ac5971e7e | 7.45 | 7 |
"""Spoken-text layer: what TTS pronounces vs what captions/readers see.
The narration string plays two roles that used to be entangled: it is the
caption/display text AND the exact characters handed to the TTS engine. The
engines (F5/Chatterbox) offer no SSML, so the only way to steer pronunciation
and pacing is to ch... | pizzato/stephen_spielbot | pipeline/tts_text.py | .py | db2f357104cff800 | 7.45 | 7 |
"""Shared UI-activity signal (issue #98).
Replaces the dedicated "ui worker" with a dynamic reservation: while the web UI
is being actively used, the render's WorkerPool holds one ComfyUI worker idle so
cover/preview jobs land on a free GPU instead of queueing behind a render. After
the UI has been idle for ``ui_idle_... | pizzato/stephen_spielbot | pipeline/ui_activity.py | .py | 612a12bc4eb03e0d | 7.45 | 7 |
"""Per-scene rendered-video history.
Every time a scene's video is re-rendered (or re-muxed after a narration change)
we keep the prior takes so the user can flip back and forth in the edit screen and
pick the best one. History lives entirely in the job's work directory — no DB
schema change, mirroring ``image_history... | pizzato/stephen_spielbot | pipeline/video_history.py | .py | d668933812923366 | 7.45 | 7 |
#!/usr/bin/env python3
"""Render `channels.yaml` into the two places the channel list is shown.
`channels.yaml` is the single source of truth — contributors add their channel
there in a pull request and touch nothing else. This script fans it out to:
* webapp/frontend/src/channels.json — bundled into the About scre... | pizzato/stephen_spielbot | scripts/gen_channels.py | .py | fe02508eff80bd84 | 7.45 | 7 |
"""Point HOME at a scratch directory before anything imports app.
app binds OUTPUT_DIR, CONFIG_FILE and LOG_DIR from ``Path.home()`` at import
time, so a test run that imports app under the real HOME reads and writes the
user's live ~/.config/video-generator/config.yaml. Individual test modules each
redirect HOME them... | pizzato/stephen_spielbot | tests/conftest.py | .py | 2b7203ca80398683 | 7.95 | 7 |
#!/usr/bin/env python3
"""Demo session gateway — single process on $PORT (8765).
Status: verified 2026-07-13 — this exact file runs the nav-trial live demo
on robium.org (Cloud Run service demo-nav-trial). Adapt FLEET_BUDGET and
the CORS origin for your deployment.
Routes:
* WebSocket upgrade (any path) -> raw byt... | robium-ai/robium | archive/live-demo/1.0.0/examples/demo_gateway.py | .py | ba4f602cb653f225 | 7.48 | 8 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.