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 |
|---|---|---|---|---|---|---|
"""The Telegram bot's inbound side: answering messages, not only sending them.
The scheduled run pushes a briefing every morning. This is the other half — asking for one
at any time and getting an answer.
Two commands, and the split between them is the whole design:
``/latest`` (or any message)
Replies instantly... | Mani5266/ai-pulse | app/delivery/bot.py | .py | ff21852e2dfc5594 | 7 | 0 |
"""Degraded-run detection.
GitHub emails the owner when a scheduled workflow fails, so an outright failure is not
silent. The failure this project could not see is the one that *succeeds*: a run that exits
zero, commits its data, deploys the site, and publishes two stories instead of five. Nobody
is watching at 02:00 ... | Mani5266/ai-pulse | app/delivery/health.py | .py | 7a4da84251479c05 | 7 | 0 |
"""Telegram delivery.
Ordering matters here and is deliberate: the briefing is **persisted before it is sent**.
Delivery is the one stage that depends on somebody else's server being up, so a failure
there must cost nothing. The briefing already exists on disk, the static site already has
it, and a failed send is retr... | Mani5266/ai-pulse | app/delivery/telegram.py | .py | ccbdffc077a74105 | 7 | 0 |
"""Prompt injection evaluation.
The number this produces is the one worth putting in a README, so it has to be earned
rather than assumed. Two rules make it honest.
**An attack is pushed through the real path.** The payload goes into an article body, and
that article goes through the same sanitisation, the same docum... | Mani5266/ai-pulse | app/evals/injection.py | .py | 063fc01640d373c3 | 7 | 0 |
"""Pipeline metrics measured against a labelled dataset and against live output.
Three of these need no human judgement and one does, and keeping the two kinds apart is
the point of this module.
**Structural metrics** are properties the pipeline must satisfy whatever anyone thinks of
the news: no story cites a source... | Mani5266/ai-pulse | app/evals/metrics.py | .py | 070173854bdcadde | 7 | 0 |
"""URL canonicalisation.
The same article reaches the pipeline under many URLs::
https://example.com/blog/model-x
https://example.com/blog/model-x/
https://www.example.com/blog/model-x?utm_source=newsletter&utm_medium=email
http://example.com/blog/model-x#section-2
https://example.com/blog/model-x... | Mani5266/ai-pulse | app/ingestion/canonical.py | .py | 5936f530c07589fb | 7 | 0 |
"""Article deduplication.
Three passes, cheapest first, because each one removes work from the next:
1. **Canonical URL** — string equality on the id derived from the canonical URL. Catches
tracking-parameter variants, ``www`` variants, trailing slashes, AMP URLs and the same
feed being re-read on a later run.
... | Mani5266/ai-pulse | app/ingestion/dedup.py | .py | 0c7204f292cfb497 | 7 | 0 |
"""Feed parsing.
Pure functions: bytes in, :class:`~app.core.models.Article` list out. No network, no
filesystem, no clock beyond the ``fetched_at`` that the caller supplies — which is what
makes this stage exhaustively testable against saved feed fixtures.
"""
from __future__ import annotations
import html
import l... | Mani5266/ai-pulse | app/ingestion/feeds.py | .py | e5a869e69fab6652 | 7 | 0 |
"""HTTP fetching with hard limits.
Every request in this project goes through :class:`SafeFetcher`. It enforces four things
that a bare ``httpx.get`` does not:
1. **SSRF validation on every hop.** The initial URL and each redirect target are
validated by :mod:`app.ingestion.urlguard` before a socket is opened.
2. ... | Mani5266/ai-pulse | app/ingestion/fetcher.py | .py | f96813f59439683e | 7 | 0 |
"""Identity and content hashes.
Two different questions, two different hashes:
* ``article_id`` answers *is this the same URL?* It is derived from the canonical URL, so
the same article under five tracking-parameter variants gets one id.
* ``content_hash`` answers *is this the same text?* It is derived from normali... | Mani5266/ai-pulse | app/ingestion/hashing.py | .py | 3737fbdd97df99b3 | 7 | 0 |
"""Article normalisation: attach canonical URL, identity and content hash.
Runs immediately after parsing, before anything tries to compare two articles. Every
later stage assumes these fields are populated.
"""
from __future__ import annotations
from collections.abc import Iterable
from app.core.models import Arti... | Mani5266/ai-pulse | app/ingestion/normalize.py | .py | 7845838a7a0d5a84 | 7 | 0 |
"""Recency filtering.
The stage that separates a news briefing from a digest of the archive.
An RSS feed does not hand you "what is new" — it hands you its current window, and the
size of that window is entirely up to the publisher. TechCrunch's twenty items are one
day; a working engineer's blog with twenty items is... | Mani5266/ai-pulse | app/ingestion/recency.py | .py | 366562ded5117979 | 7 | 0 |
"""Ingestion orchestration.
One dead feed must never end a run, so every failure is captured as a
:class:`~app.core.models.FeedResult` value rather than raised. The run statistics in P10
are built from exactly these records.
"""
from __future__ import annotations
import logging
import time
from collections.abc impor... | Mani5266/ai-pulse | app/ingestion/runner.py | .py | 55f8b750f010d3d4 | 7 | 0 |
"""Source registry loading.
Sources live in ``config/sources.yaml`` rather than in code so that adding a feed is a
data change, reviewable as a one-line diff, with no import cycle and no redeploy.
"""
from __future__ import annotations
from collections.abc import Iterable
from pathlib import Path
from typing import ... | Mani5266/ai-pulse | app/ingestion/sources.py | .py | cddea98e56de44d7 | 7 | 0 |
"""SSRF guard.
Feed URLs come from a registry file, but redirects come from the open internet, so every
hop is validated before a socket is opened.
The check is performed on the **resolved address**, not on the hostname. A hostname
allowlist or a regular expression on the host is not a defence: an attacker-controlled... | Mani5266/ai-pulse | app/ingestion/urlguard.py | .py | 79c279cd7c360cc7 | 7 | 0 |
"""Event clustering: articles in, events out.
This is the stage the whole product rests on. Deduplication removes copies of one
*article*; clustering groups different articles about one *development*. Four outlets
covering a model release become one event with four sources — which is both what the
briefing should say ... | Mani5266/ai-pulse | app/intelligence/clustering.py | .py | 3f81cc4e80d84a41 | 7 | 0 |
"""Entity extraction.
Clustering needs to know what an article is *about* before it can decide that two
articles are about the same thing. Title similarity alone is not enough: "OpenAI releases
GPT-X" and "GPT-X is now available to developers" share almost no characters but are
obviously one event, while "Google launc... | Mani5266/ai-pulse | app/intelligence/entities.py | .py | 2107989631e7384e | 7 | 0 |
"""Candidate pairs for duplicate adjudication.
Clustering in P3 is precision-tuned and under-clusters on purpose: two outlets describing
one development in genuinely different words stay two events unless they name the same
model, lab or version. `PLAN.md` §2.9 argues that trade — a false merge puts two unrelated
stor... | Mani5266/ai-pulse | app/intelligence/pairing.py | .py | f9079e501295a832 | 7 | 0 |
"""Title similarity.
Character trigrams with the Dice coefficient. No embeddings, no model, no vector
database — and that is a deliberate choice, not a shortcut:
* It is deterministic, so a test written today still passes next year.
* It is fast enough that all-pairs comparison over a day's articles is milliseconds.
... | Mani5266/ai-pulse | app/intelligence/similarity.py | .py | ae6e571b0a898ffe | 7 | 0 |
"""Event timelines: how a story changed, day by day.
This is the feature that separates an intelligence system from a newsletter, and it needs
no new data. Every run writes a full snapshot of the events it touched, append-only, one
file per day — so an event's history is already sitting in the repository and can be
re... | Mani5266/ai-pulse | app/intelligence/timeline.py | .py | b2cb89c8ec448479 | 7 | 0 |
"""Claim verification.
The stage that makes "evidence-backed" a fact about the output rather than a word in the
README.
**The split.** The model extracts claims and says which documents assert each one. This
module assigns the label, in code, by counting *independent sources*. That division is
deliberate and is the s... | Mani5266/ai-pulse | app/intelligence/verification.py | .py | b7ba03828b420a12 | 7 | 0 |
"""Run the bot: answer messages until stopped.
python -m app.jobs.serve_bot
Long-polling, so no public endpoint, no webhook, no server. The process holds one idle
connection to Telegram and wakes when a message arrives.
This is the interactive half of the product. The scheduled run pushes a briefing each
morning... | Mani5266/ai-pulse | app/jobs/serve_bot.py | .py | 05bbb4cee16c610d | 7 | 0 |
"""The model stage of the pipeline.
Everything the model is asked to do, in one place, so the total cost of a run is countable
by reading one file:
===================== ============================ ==============================
Call Purpose Count per run
===================== =... | Mani5266/ai-pulse | app/llm/analysis.py | .py | ab7296d8f9b27e1a | 7 | 0 |
"""A chain of free-tier providers, tried in order.
Every free allowance runs out. Groq's is 200,000 tokens a day, which is about five runs;
Cerebras and OpenRouter have their own. One key is therefore a single point of failure with
a known failure time, and the pipeline's answer until now was to degrade to the
determi... | Mani5266/ai-pulse | app/llm/chain.py | .py | f542d2b3021600a4 | 7 | 0 |
"""Prompt construction and untrusted-content framing.
This module is the trust boundary. Everything above it is data the pipeline controls;
everything passed through :func:`wrap_documents` is text harvested from the open internet
and must be treated as hostile.
The defence has three parts, and none of them is "ask th... | Mani5266/ai-pulse | app/llm/prompts.py | .py | 528ae7dd26d06acd | 7 | 0 |
"""Schemas for every model response.
The rule this module enforces: **no model output reaches the rest of the application as
free text**. Every call declares a Pydantic model, the response is validated against it,
and a response that does not validate is discarded rather than parsed leniently.
That is not defensive s... | Mani5266/ai-pulse | app/llm/schemas.py | .py | 27bff5cbdfc8db60 | 7 | 0 |
"""Personal relevance profile.
Taste lives in ``config/profile.yaml`` rather than in code, for the same reason the source
registry does: changing what you care about should be a one-line data diff, not a code
change, and it should be visible to anyone reading the repository.
"""
from __future__ import annotations
im... | Mani5266/ai-pulse | app/ranking/profile.py | .py | 0159d0f769c4c801 | 7 | 0 |
from __future__ import annotations
import ipaddress
import logging
import time
import unicodedata
from collections.abc import Callable
from concurrent.futures import ThreadPoolExecutor
from typing import Any
from urllib.parse import urlsplit
from .controller_common import (
NETWORK_SCAN_MAX_THREADS,
NETWORK_S... | SurreptitiousFabric/omarchy-sonarchy | sonarchy_backend/controller_discovery.py | .py | 4024d7ec0d50bb78 | 7 | 0 |
from __future__ import annotations
import logging
import time
from collections.abc import Callable, Iterable
from typing import Any
from .controller_common import (
TOPOLOGY_QUERY_TIMEOUT_SEC,
TOPOLOGY_SETTLE_ATTEMPTS,
TOPOLOGY_SETTLE_INTERVAL_SEC,
ControllerError,
)
LOG = logging.getLogger(__name__)... | SurreptitiousFabric/omarchy-sonarchy | sonarchy_backend/controller_topology.py | .py | cfca43fbc06b95fe | 7 | 0 |
from __future__ import annotations
from typing import Any
import requests
from .common import clean, safe_call
LINE_IN_QUERY_TIMEOUT_SEC = 1.5
def tv_autoplay_enabled(speaker: Any) -> bool | None:
"""Project TV Autoplay support and state without guessing from a model name."""
response = safe_call(
... | SurreptitiousFabric/omarchy-sonarchy | sonarchy_backend/domains/capabilities.py | .py | 54b9de87e06c4782 | 7 | 0 |
"""Turn low-level SoCo/UPnP failures into short recovery instructions."""
from __future__ import annotations
import re
from typing import Any
_UPNP_CODE = re.compile(r"(?:UPnP Error|error(?:_code)?[=: ]+)\s*['\"]?(\d{3})", re.I)
_TRAILING_HOST = re.compile(r"\s+from\s+(?:\d{1,3}\.){3}\d{1,3}(?::\d+)?\s*$", re.I)
_U... | SurreptitiousFabric/omarchy-sonarchy | sonarchy_errors.py | .py | 16a95c90fcdcdf4d | 7 | 0 |
from __future__ import annotations
from collections.abc import Mapping, Sequence
from dataclasses import dataclass
from datetime import UTC, datetime
from application.evaluations.risk_authority_gate import (
RiskAuthorityGateDecision,
RiskAuthorityGateEvidence,
)
from application.observability.ai_observabilit... | sponge-b0b/Polaris | application/evaluations/contracts.py | .py | 4f308b70b9217f8a | 7.35 | 4 |
from __future__ import annotations
from collections.abc import Sequence
from dataclasses import dataclass
from application.evaluations.contracts import EvaluationResultBundle
from core.storage.persistence.evaluation import (
EvaluationCaseRecord,
EvaluationDatasetRecord,
EvaluationPersistenceRepository,
... | sponge-b0b/Polaris | application/evaluations/evaluation_result_service.py | .py | 428b77eeda09d89d | 7.35 | 4 |
"""Durable reconstruction boundary for Baseline runtime provenance."""
from typing import Protocol
from domain.governed_execution_evidence import BaselineRuntimeEvidence
class BaselineRuntimeEvidenceNotFoundError(LookupError):
"""Raised when canonical Baseline runtime provenance is unavailable."""
class Basel... | sponge-b0b/Polaris | application/governance/baseline_runtime_evidence.py | .py | 7a04e86043901754 | 7.35 | 4 |
from __future__ import annotations
from collections.abc import Mapping
from typing import Any
from domain.authority import RISK_AUTHORITY_METADATA_KEY, RiskAuthorityContract
_AUTHORITY_METADATA_LABELS: Mapping[str, str] = {
"risk_tier": "authority_risk_tier",
"authority_effect": "authority_effect",
"cont... | sponge-b0b/Polaris | application/observability/risk_authority.py | .py | c143f65af494d44f | 7.35 | 4 |
"""Remap region codes unsupported by the new prayer-time provider.
Revision ID: 20260826_02
Revises: 20260826_01
"""
from collections.abc import Sequence
import sqlalchemy as sa
from alembic import op
revision: str = "20260826_02"
down_revision: str | None = "20260826_01"
branch_labels: str | Sequence[str] | None ... | turgunovjasur/Namoz-vaqti | alembic/versions/20260826_02_remap_unsupported_regions.py | .py | dd3d2eef085fb98c | 7 | 0 |
"""Enable daily notifications for all existing users.
Revision ID: 20260826_04
Revises: 20260826_03
"""
from collections.abc import Sequence
import sqlalchemy as sa
from alembic import op
revision: str = "20260826_04"
down_revision: str | None = "20260826_03"
branch_labels: str | Sequence[str] | None = None
depend... | turgunovjasur/Namoz-vaqti | alembic/versions/20260826_04_enable_daily_notifications.py | .py | ecedf5553ea22ecb | 7 | 0 |
"""Dependency-inversion ports owned by the application layer."""
from datetime import date
from typing import Protocol
from namoz_bot.domain.models import (
DeliveryStatus,
DeliveryType,
OffsetAction,
PrayerKey,
PrayerSchedule,
UserSubscription,
)
class PrayerScheduleProvider(Protocol):
... | turgunovjasur/Namoz-vaqti | src/namoz_bot/application/ports.py | .py | c4891ddb8d6d8cc2 | 7 | 0 |
"""User onboarding and preference use cases."""
from dataclasses import dataclass
from namoz_bot.application.ports import SubscriptionRepository
from namoz_bot.domain.errors import SubscriptionNotFoundError
from namoz_bot.domain.models import OffsetAction, PrayerKey, UserSubscription
from namoz_bot.domain.regions imp... | turgunovjasur/Namoz-vaqti | src/namoz_bot/application/subscriptions.py | .py | 5bc5608cca057d3d | 7 | 0 |
"""Immutable business entities shared by all adapters."""
from dataclasses import dataclass, field, replace
from datetime import date, datetime
from enum import StrEnum
from typing import Literal
from namoz_bot.domain.errors import ScheduleValidationError
PrayerKey = Literal["bomdod", "quyosh", "peshin", "asr", "sho... | turgunovjasur/Namoz-vaqti | src/namoz_bot/domain/models.py | .py | e177949c27f51a82 | 7 | 0 |
"""Rate-limited Telegram message-sending adapter."""
import asyncio
import time
from collections.abc import Awaitable, Callable
from typing import Protocol
from aiogram import Bot
from aiogram.exceptions import (
TelegramBadRequest,
TelegramForbiddenError,
TelegramNetworkError,
TelegramRetryAfter,
... | turgunovjasur/Namoz-vaqti | src/namoz_bot/infrastructure/telegram.py | .py | 6cf50a11141cbaf9 | 7 | 0 |
"""Reusable Telegram keyboards."""
from aiogram.types import (
InlineKeyboardButton,
InlineKeyboardMarkup,
KeyboardButton,
ReplyKeyboardMarkup,
)
from namoz_bot.domain.models import PrayerKey
from namoz_bot.domain.regions import list_region_groups, list_regions
TODAY_LABEL = "📅 Bugungi jadval"
REGIO... | turgunovjasur/Namoz-vaqti | src/namoz_bot/presentation/keyboards.py | .py | 7b73369d6e98c989 | 7 | 0 |
"""Request-scoped service injection over short-transaction repositories."""
import logging
from collections.abc import Awaitable, Callable
from datetime import datetime
from typing import Any
from zoneinfo import ZoneInfo
from aiogram import BaseMiddleware
from aiogram.types import TelegramObject
from sqlalchemy.ext.... | turgunovjasur/Namoz-vaqti | src/namoz_bot/presentation/middleware.py | .py | 946bbd7bc06b328c | 7 | 0 |
"""Daily APScheduler wiring."""
from collections.abc import Awaitable, Callable
from datetime import date, datetime, time, timedelta
from zoneinfo import ZoneInfo
from apscheduler.schedulers.asyncio import AsyncIOScheduler
from apscheduler.triggers.cron import CronTrigger
from namoz_bot.application.broadcasting impo... | turgunovjasur/Namoz-vaqti | src/namoz_bot/scheduler.py | .py | 1c3a16f4122c6a77 | 7 | 0 |
import io
import os
import tempfile
import contextlib
import zstandard as zstd
LOG_COMPRESSION_LEVEL = 10 # little benefit up to level 15. level ~17 is a small step change
class CallbackReader:
"""Wraps a file, but overrides the read method to also
call a callback function with the number of bytes read so far.""... | gotoded/sunnypilot | common/file_helpers.py | .py | 2bc174323e1b3cca | 7 | 0 |
"""Utilities for reading real time clocks and keeping soft real time constraints."""
import gc
import os
import time
from setproctitle import getproctitle
from openpilot.common.util import MovingAverage
from openpilot.system.hardware import PC
# time step for each process
DT_CTRL = 0.01 # controlsd
DT_MDL = 0.1 #... | gotoded/sunnypilot | common/realtime.py | .py | 92b418b21db0da39 | 7 | 0 |
import signal
class TimeoutException(Exception):
pass
class Timeout:
"""
Timeout context manager.
For example this code will raise a TimeoutException:
with Timeout(seconds=5, error_msg="Sleep was too long"):
time.sleep(10)
"""
def __init__(self, seconds, error_msg=None):
if error_msg is None:
... | gotoded/sunnypilot | common/timeout.py | .py | 862433064831fb47 | 7 | 0 |
import itertools
import numpy as np
from dataclasses import dataclass
import openpilot.common.transformations.orientation as orient
## -- hardcoded hardware params --
@dataclass(frozen=True)
class CameraConfig:
width: int
height: int
focal_length: float
@property
def size(self):
return (self.width, sel... | gotoded/sunnypilot | common/transformations/camera.py | .py | 1538a95c0daa0f80 | 7 | 0 |
"""
atlas_run.py
============
II.7 Atlas-scale reduction pipeline (Session 2, unblocked).
Uses Kang et al. 2025 (Nat Methods, doi:10.1038/s41592-025-02857-2) Supplementary
Table S11 as the per-cell score matrix — the atlas's own author-computed values for
CytoTRACE 2, CytoTRACE 1 (gene counts), SCENT (CCAT), SCENT (SR... | IskakovDamir/potency-ownbaseline | experiments/atlas_run/atlas_run.py | .py | 1a39e91c991631e5 | 7.35 | 4 |
"""
h6_partial_rho.py
=================
H6 (difficulty-controlled separation): partial-rho analysis per prereg
2026-07-17-hod6-sign-rule-and-separation-prereg.md.
Computes Spearman partial correlation
partial-rho(tau_SR_d, tau_gc_d | tau_CT2_d)
partial-rho(tau_CCAT_d, tau_gc_d | tau_CT2_d)
under both kernels (... | IskakovDamir/potency-ownbaseline | experiments/atlas_run/h6_partial_rho.py | .py | e5d6e4d47b554e2e | 7.35 | 4 |
"""
pi_gate_recompute.py
====================
PI-gate 2026-07-17: independently recompute CT2 median Δ vs gene_counts on the
23 H1 rows, under both kernels, from the live per_dataset_results.json.
Purpose: resolve the same-document contradiction in the dossier (§4 vs §8):
§4 lines 108-109: scipy +0.042, Kang +0.088
... | IskakovDamir/potency-ownbaseline | experiments/atlas_run/pi_gate_recompute.py | .py | fd7b7cf4c3e73f53 | 7.35 | 4 |
"""
wdm_coverage_14rows.py
======================
PI-gate 2026-07-17, mandate 3.
Current wdm validation covers ONE dataset (Retinal neurons 10x) against Kang's
S12. Kang's S12 has 13 Test-cohort rows (mandate said 14; audit shows 13 non-
median rows in Test cohort — one row appears absent, likely reflecting Kang's
own... | IskakovDamir/potency-ownbaseline | experiments/atlas_run/wdm_coverage_14rows.py | .py | f403b2755d064516 | 7.35 | 4 |
"""
CytoTRACE 2 own-baseline probe (Question A, single dataset - human cord blood CITE-seq).
Purpose
-------
Second dataset in the pre-registered probe (SSOT: CT2 own-baseline prereg 2026-07-12).
Same protocol as pancreas (probe_ct2_ownbaseline_pancreas.py).
Compares within-dataset weighted Kendall tau of two potency... | IskakovDamir/potency-ownbaseline | experiments/ct2_probe/probe_ct2_ownbaseline_cordblood.py | .py | 5833e5b71306b326 | 7.35 | 4 |
"""
CytoTRACE 2 own-baseline probe (Question A, dataset 3 - Paul et al. 2015 mouse
hematopoiesis, Drop-seq).
Purpose
-------
Third dataset in the pre-registered multi-dataset probe. Same protocol as
pancreas + cord blood. Complements the two vignette datasets with a
Drop-seq dataset spanning committed progenitors -> m... | IskakovDamir/potency-ownbaseline | experiments/ct2_probe/probe_ct2_ownbaseline_paul15.py | .py | a1fc48535eab43fd | 7.35 | 4 |
"""
test_reproducibility.py — GATE 1 reproducibility check for StemSC.
Runs the Python StemSC implementation on Zhao et al.'s own example data
(GSE85066 H7 hESC RPKM matrix, shipped with the R package as data/example.RData)
and confirms it reproduces the ESC-ceiling behaviour reported in the paper
(median ESC StemSC =... | IskakovDamir/potency-ownbaseline | experiments/stemsc/test_reproducibility.py | .py | 7cb7e06c46315970 | 7.85 | 4 |
"""
L2 EXPANSION scoring pipeline — ssGSEA + Cox + library-size-adjusted Venet null.
PI convention (locked 2026-07-20):
Signature = the gene list the paper APPLIES to the bulk cohort (published supp table).
Parse verbatim, freeze, no reselection.
Score = ssGSEA enrichment of the signature per bulk sample (unifor... | IskakovDamir/potency-ownbaseline | experiments/track1/code/l2_ssgsea_score.py | .py | a7892eabc91407da | 7.35 | 4 |
"""Configuration for the authenticated DataPulse buyer API."""
from __future__ import annotations
import os
from dataclasses import dataclass
from pathlib import Path
# Public metadata only. These values intentionally mirror the existing
# runtime fallbacks and caps without reading an effective environment.
PAGINAT... | r3dz4r/datapulse-my | api/config.py | .py | c3851a02a7693279 | 7 | 0 |
"""Public, immutable metadata for the existing buyer API routes.
This deliberately describes only literal route shapes and public query inputs;
it is not an OpenAPI document and it never reads process configuration.
"""
from __future__ import annotations
from dataclasses import dataclass
@dataclass(frozen=True)
cl... | r3dz4r/datapulse-my | api/public_contract.py | .py | 0d01310127c4406a | 7 | 0 |
#!/usr/bin/env python3
"""Render the public health-methodology HTML page with Pandoc."""
from __future__ import annotations
import os
import re
import shutil
import subprocess
import sys
import tempfile
from pathlib import Path
try: # Support both ``python scripts/...`` and package imports in tests.
from script... | r3dz4r/datapulse-my | scripts/gen_health_methodology_html.py | .py | 704c6101d0cf9f4a | 7 | 0 |
#!/usr/bin/env python3
"""Generate canonical sitemap and marker-owned public discovery blocks."""
from __future__ import annotations
import argparse
import re
import sys
from pathlib import Path
from xml.sax.saxutils import escape
sys.path.insert(0, str(Path(__file__).resolve().parents[1]))
from scripts.public_surf... | r3dz4r/datapulse-my | scripts/gen_public_discovery.py | .py | 039508479c76bece | 7 | 0 |
#!/usr/bin/env python3
"""Inject the canonical site navigation into the public HTML pages."""
from __future__ import annotations
import argparse
import sys
from html.parser import HTMLParser
from pathlib import Path
SCRIPT_ROOT = Path(__file__).resolve().parents[1]
if str(SCRIPT_ROOT) not in sys.path:
sys.path.i... | r3dz4r/datapulse-my | scripts/gen_site_nav.py | .py | 6769862debda9d06 | 7 | 0 |
#!/usr/bin/env python3
"""Create a protected Ed25519 probe key and append its public registry row."""
from __future__ import annotations
import argparse, base64, hashlib, json, os
from datetime import datetime, timedelta, timezone
from pathlib import Path
from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed... | r3dz4r/datapulse-my | scripts/init_keys.py | .py | 823a67b71c2c2cc4 | 7 | 0 |
"""Regenerate gRPC stubs from ``proto/scheduler.proto``.
Run from the package root:
cd packages/z4j-scheduler
python scripts/regen_proto.py
This script:
1. Compiles the .proto file via ``grpc_tools.protoc`` into both:
- ``packages/z4j-scheduler/src/z4j_scheduler/proto/``
- ``packages/z4j/backend/src/z... | z4jdev/z4j-scheduler | scripts/regen_proto.py | .py | 157e3f9d36e356b3 | 7 | 0 |
"""Shared runtime state for the FastAPI operational endpoints.
Constructed once at startup by the SchedulerApp lifespan and stuck
on ``app.state.scheduler_state``. Each endpoint reads fields from
it via FastAPI ``Depends()`` to render /health, /ready, /info.
Kept deliberately simple - just a dataclass. The endpoints ... | z4jdev/z4j-scheduler | src/z4j_scheduler/api/_state.py | .py | 19c63a197b1eab35 | 7 | 0 |
"""Flask integration for declarative schedule reconciliation.
Flask's startup hooks differ enough from Django/FastAPI that the
helper takes a different shape: a function the operator calls to
register a CLI command + an explicit one-shot reconciler the
operator invokes from their app factory.
There is no ``before_fir... | z4jdev/z4j-scheduler | src/z4j_scheduler/declarative/frameworks/flask.py | .py | 17a5968d7fa92daa | 7 | 0 |
"""Django ``AppConfig`` for the z4j-scheduler integration.
The minimal contract: register the app so ``manage.py
z4j_schedules ...`` is discoverable. The optional contract:
auto-reconcile on web-worker startup when
``settings.Z4J_SCHEDULES_AUTO_RECONCILE`` is True.
Auto-reconcile is OFF by default. Most operators pre... | z4jdev/z4j-scheduler | src/z4j_scheduler/django_app/apps.py | .py | 2058c65d018d554c | 7 | 0 |
"""REST client + shared dataclass for the exporters.
Fetches the schedule set from brain via
``GET /api/v1/projects/{slug}/schedules`` and converts each row to
the :class:`ExportedSchedule` shape the per-target renderers
consume.
Symmetric with :class:`z4j_scheduler.importers._core.BrainImportClient`
but read-side. W... | z4jdev/z4j-scheduler | src/z4j_scheduler/exporters/_client.py | .py | 71b912e11ec13d36 | 7 | 0 |
"""Render exported schedules as a celery-beat ``beat_schedule`` dict.
Output is a Python module the operator can drop into their Celery
app config:
# generated by `z4j-scheduler export --to celery ...`
from celery.schedules import crontab
from datetime import timedelta
beat_schedule = {
"ever... | z4jdev/z4j-scheduler | src/z4j_scheduler/exporters/celery.py | .py | 2b2f625f576fffa4 | 7 | 0 |
"""Render exported schedules as a system crontab file.
Output is a 5-field crontab. Because z4j tasks are not shell
commands, the export needs an operator-supplied wrapper script
that invokes the named task. The header comment names the
expected wrapper convention so the operator can install one.
Example:
# Gene... | z4jdev/z4j-scheduler | src/z4j_scheduler/exporters/cron.py | .py | 64e23016c3a7fb39 | 7 | 0 |
"""Leader election - ensures only one scheduler instance dispatches per project.
Two implementations:
- :class:`SingleInstanceLeaderGate` - always returns True. The right
choice for solo deployments where exactly one scheduler process
ever runs. Adds zero infrastructure dependencies.
- :class:`~z4j_scheduler.lead... | z4jdev/z4j-scheduler | src/z4j_scheduler/leader/__init__.py | .py | 82d91c2aabdf1b52 | 7 | 0 |
"""structlog configuration matching z4j-brain's setup.
Configured once at process startup based on
``Z4J_SCHEDULER_LOG_LEVEL`` and ``Z4J_SCHEDULER_LOG_JSON``.
In production (``log_json=True``), emits one JSON object per log
entry with ISO timestamp, level, logger name, event message, and
any bound contextual fields. ... | z4jdev/z4j-scheduler | src/z4j_scheduler/observability/logging.py | .py | c37c71da6a38e48b | 7 | 0 |
#!/usr/bin/env python3
"""Initialize a manage-math-research-program repository.
This script creates only project-management files. It never creates any
problem-level artifacts owned by $rigorous-open-math-research.
"""
from __future__ import annotations
import argparse
import json
import re
import shutil
from dateti... | xsoc1/math-research-dsh | skills/manage-math-research-program/scripts/init_project.py | .py | 30d731525e163ebc | 7.24 | 2 |
#!/usr/bin/env python3
"""Generate a Lean scaffold and its registration records.
This script automates the "scaffold every new result" rule (manage workflow 8d)
and the "proof submission audit" record (workflow 8e). It creates:
1. `lean-proof/SL/<slug>.lean` - a Lean scaffold with a `-- SCAFFOLD` header.
2. An en... | xsoc1/math-research-dsh | skills/manage-math-research-program/scripts/scaffold_result.py | .py | 2593671168e55cdd | 7.24 | 2 |
#!/usr/bin/env python3
"""Maintain the human-readable research map for a project.
The research map is a living, survey-style document (`research_map.md`) that
collects every route/method tried, intermediate results, unexpected findings,
failures and their reasons, tools, open directions, an avoid list, and human /
oth... | xsoc1/math-research-dsh | skills/manage-math-research-program/scripts/update_research_map.py | .py | 5c53d3acdfc30634 | 7.24 | 2 |
#!/usr/bin/env python3
"""Smoke test for lake_build_guard.py.
Verifies that the guard:
1. allows a first build check;
2. refuses when a fresh lock exists;
3. releases the lock;
4. refuses when too many recent build attempts are logged;
5. allows again after clearing state.
"""
from __future__ import annotat... | xsoc1/math-research-dsh | tests/smoke_lake_build_guard.py | .py | 3716ac464a8a7e14 | 7.74 | 2 |
#!/usr/bin/env python3
"""Smoke test: the pipeline gate must ignore nested git repositories.
A project may contain a cloned plugin repo (e.g. `_xsoc1_work/`) whose test
fixtures intentionally contain failing handoffs/whiteboards. The gate must not
validate those as part of the parent project.
"""
from __future__ impo... | xsoc1/math-research-dsh | tests/smoke_nested_repo.py | .py | b16cc9f0015f74f0 | 7.74 | 2 |
#!/usr/bin/env python3
"""Smoke test for scripts/check_version_bump.py.
Creates a throwaway git repo and verifies:
1. a commit changing skills/ without package.json fails the gate;
2. a commit changing skills/ together with package.json passes the gate.
"""
from __future__ import annotations
import pathlib
impor... | xsoc1/math-research-dsh | tests/smoke_version_bump.py | .py | 18a341f77bd4d95c | 7.74 | 2 |
import os
from playwright.sync_api import Page, expect
# Muhitni TEST_ENV environment variable tanlaydi: "dev" (default) yoki "prod".
# Telegram bot (tg_bot_runner.py) "start prod" / "start dev" bilan aynan shu
# env'ni beradi — endi faylni qo'lda tahrirlash SHART EMAS. Berilmasa yoki
# noto'g'ri qiymatда DEV (sm24) ... | turgunovjasur/Smartup24 | flows/flow_authorization.py | .py | e3ef8a96e05dc657 | 7 | 0 |
"""Smartup24 formalari ochilish smoke testi.
Maqsad: har bir navbar bo'limini (Модератор / Поставщик / Клиент menyusidagi
formalar) ketma-ket ochib, forma to'g'ri yuklanganini (aktiv sarlavha paydo
bo'lishi) va ochilishda xato chiqmasligini tekshirish. Bitta forma yiqilsa ham
to'xtamaydi — keyingisini ochadi va oxirid... | turgunovjasur/Smartup24 | tests/test_forms_smoke.py | .py | 1cf36d959407a908 | 7.5 | 0 |
import random
import allure
from playwright.sync_api import Page
from flows.flow_authorization import authorization
from flows.flow_navbar import flow_navigate
from utils.base_page import BasePage
def run_client_user(page: Page, code) -> None:
"""Group A: client-{code} ning Просмотр formasida Пользователи bo'li... | turgunovjasur/Smartup24 | tests/test_group_a/test_client_user.py | .py | fa6bb68f1a9b8ab9 | 7.5 | 0 |
import re
import allure
from playwright.sync_api import Page, expect
from flows.flow_authorization import authorization
from flows.flow_navbar import flow_navigate
from utils.base_page import BasePage
def run_cooperation(page: Page, code) -> None:
"""Group A: supplier-{code} dan client-{code} ga hamkorlik so'ro... | turgunovjasur/Smartup24 | tests/test_group_a/test_cooperation.py | .py | 1dce0a2ef8ea21ea | 7.5 | 0 |
from datetime import datetime, timedelta
import allure
from playwright.sync_api import Page, expect
from flows.flow_authorization import COMPANY_CODE, authorization
from flows.flow_navbar import flow_navigate
from utils.base_page import BasePage
def _set_qty(page: Page, value: str = "2", *, attempts: int = 5) -> No... | turgunovjasur/Smartup24 | tests/test_group_a/test_order.py | .py | 0567b62a920240a1 | 7.5 | 0 |
import allure
from playwright.sync_api import Page
from flows.flow_authorization import COMPANY_CODE, authorization
from flows.flow_navbar import flow_navigate
from utils.base_page import BasePage
# Order hayot sikli — "Изменить статус" menyusida har safar joriy statusdan
# keyingi bosqich tanlanadi; order run_order ... | turgunovjasur/Smartup24 | tests/test_group_a/test_order_status_change.py | .py | e681e47f736b24d4 | 7.5 | 0 |
import random
import allure
from playwright.sync_api import Page
from flows.flow_authorization import authorization
from flows.flow_navbar import flow_navigate
from utils.base_page import BasePage
def run_supplier_user(page: Page, code) -> None:
"""Group A: supplier-{code} ning Просмотр formasida Пользователи b... | turgunovjasur/Smartup24 | tests/test_group_a/test_supplier_user.py | .py | 520cc13a81afc1e1 | 7.5 | 0 |
"""Настройка (Модератор → Главное → Настройка, biruni moderator/setting) —
kompaniya sozlamalari sahifasi smoke testi.
Loyiha-playwright bilan tasdiqlangan:
- URL .../sb/sbr/moderator/setting, heading "Настройка".
- Yagona sozlamalar formasi: ~11 textbox + 3 switch + "Сохранить".
- Bu KOMPANIYA darajasidagi GLOBAL soz... | turgunovjasur/Smartup24 | tests/test_main/test_settings.py | .py | 3660390e80b64f40 | 7.5 | 0 |
"""Перевод строки таблицы (Модератор → Главное, biruni md/table_translate_list) —
jadval yozuvlari tarjimasi testi.
MCP bilan real DOM'da tasdiqlangan:
Bu modul CREATE-CRUD EMAS — oldindan belgilangan tizim jadvallarining ustun/yozuv
nomlarini tillarga (uz/ru/en) tarjima qilish. "Создать" YO'Q.
Oqim:
- table_trans... | turgunovjasur/Smartup24 | tests/test_main/test_table_translate.py | .py | 9afd42c00576d007 | 7.5 | 0 |
"""S3 (MinIO) configuration.
Config comes from environment variables only. No credentials in code, in
defaults, or in tests: anything in an environment variable lands in
`docker inspect` and in the pod spec.
"""
from __future__ import annotations
import os
from dataclasses import dataclass
_REQUIRED_VARS = (
"K... | WawRepo/KaBOM | kabom/config.py | .py | 750eb8b3f64eb4a3 | 7 | 0 |
"""Parsing for CycloneDX SBOM documents.
KaBOM only ever reads CycloneDX JSON produced by Syft (HOME-224). This module
has no opinion about vulnerabilities or policy — it extracts exactly what the
search feature (HOME-231) and the freshness banner (HOME-232) need: the
subject, the generation timestamp, and the compone... | WawRepo/KaBOM | kabom/cyclonedx.py | .py | f2e2861fc157ee0f | 7 | 0 |
"""SQLite schema setup and transactional ingest.
Plain `sqlite3` and `schema.sql` — no ORM, no migrations. If the schema
changes, drop the database file and re-ingest; the source of truth is S3
(kabom.ingest) and a full rebuild takes seconds.
This module writes the already-parsed output of kabom.ingest.ingest_all int... | WawRepo/KaBOM | kabom/db.py | .py | a7b5a9f8a5ff6ce5 | 7 | 0 |
"""Fetch every SBOM out of the configured bucket and parse it.
This is the read half of HOME-229: list the bucket, fetch each object, parse
it as CycloneDX, and hand back everything that parsed plus a countable,
named record of everything that didn't. No database yet — that's HOME-230.
This module ends with parsed obj... | WawRepo/KaBOM | kabom/ingest.py | .py | ef46041f7a6ba2ec | 7 | 0 |
"""Thin, read-only S3 (MinIO) access.
KaBOM never writes to the bucket, ever. This module only lists and fetches objects.
"""
from __future__ import annotations
from collections.abc import Iterator
import boto3
from botocore.exceptions import BotoCoreError, ClientError
from kabom.config import S3Config
class S3U... | WawRepo/KaBOM | kabom/s3_client.py | .py | 9ac76c5df8153b96 | 7 | 0 |
#!/usr/bin/env python3
"""Seed the dev/test MinIO bucket used by docker-compose.yml with sample
CycloneDX files.
This is dev/test tooling only — see docker-compose.yml's `seed` service. It
runs against the compose-local MinIO container, never your real MinIO,
and it is never copied into the production image (the Docke... | WawRepo/KaBOM | scripts/seed_minio.py | .py | 419fa22482225513 | 7 | 0 |
"""配置中心:端口、上游、数据目录、类别(全部可被界面设置覆盖)。
优先级(FR-15):settings.json(界面设置) > 环境变量 > 内置默认值。
"""
import json
import os
import tempfile
from pathlib import Path
def _windows_data_dir(localappdata=None) -> str:
"""Windows 数据目录字符串:%LOCALAPPDATA%\\llm-sanitizer(纯函数,可测)。"""
base = localappdata or os.environ.get("LOCALAPPDA... | JunyuZhan/llm-sanitizer | llm_sanitizer/config.py | .py | 73fce76f3a3c2514 | 7 | 0 |
"""桌面轻包(v0.5,可选特性):原生窗口打开控制台,退出时关闭网关/看板。
设计(见 docs/开发文档.md ADR-12/14):
- 界面层与内核解耦:网关/看板仍是纯标准库,本模块只是"打开浏览器窗口的壳"
- **可选依赖,ADR-1 核心零依赖不受影响**:
pip install llm-sanitizer-gateway[desktop] # pywebview
- 定位:小白用户免浏览器标签页——"一键打开看数据,关窗即走";
v0.5.3 起:后台已有服务(开机自启/手动 start)时直接开窗看数据,
不重复起服务,关闭窗口也不影响后台服务(那是 launchd 托管的)。
"""
... | JunyuZhan/llm-sanitizer | llm_sanitizer/desktop.py | .py | 27f456488b5b8bfc | 7 | 0 |
"""事件存储:JSONL 追加写,线程安全。
设计约束:事件只记录占位符/类别/时间/请求路径,**绝不记录明文**。
明文只在 map.json 中,供还原使用。
**累计统计(FR-5 修订)**:`stats.json`(同目录,权限 600)持久化 per-category
计数器——看板"累计脱敏数"不随事件文件轮转/超尾而倒退,也不随网关重启归零。
事件文件本身只做最近事件的展示与轮转。
"""
import json
import os
import tempfile
import threading
import time
from collections import deque
class Event... | JunyuZhan/llm-sanitizer | llm_sanitizer/events.py | .py | c026db113cc1b642 | 7 | 0 |
"""图片 OCR 脱敏(v0.4,可选依赖特性)。
定位:与 docx/xlsx/pdf 同一"辅助链路"——先把身份证照片、合同扫描件等
图片里的文字 OCR 出来并脱敏,再交给 Agent,避免图片中的明文出网。
**为什么是可选特性(ADR-1 承诺不变)**:OCR 无法用纯标准库实现,需要
第三方库 + 系统二进制,故设计为:
pip install llm-sanitizer-gateway[ocr]
# 还需系统级 tesseract:brew install tesseract tesseract-lang
# / apt install tesseract-ocr tesserac... | JunyuZhan/llm-sanitizer | llm_sanitizer/ocr.py | .py | 1d0915bca12cc5ae | 7 | 0 |
"""看板前端完整性测试(v0.6.1):PAGE 模板的 JS 不得出现语法级错误。
背景:v0.5 引入 `def loadSettings() {`(Python 关键字混入 JS),导致整个
<script> 块语法错误、一行不跑——统计永远 0、Agent 永远"检测中…",且
curl/API 测试全部正常,只有浏览器渲染才暴露。本测试做静态守卫:
- PAGE 内不允许行首 `def `(JS 没有 def 关键字)
- 关键函数必须存在且数量正确
- 可选:node --check 深度校验(CI 有 node;本地缺失时跳过)
"""
import os
import re
import shutil
impo... | JunyuZhan/llm-sanitizer | tests/test_dashboard.py | .py | 93994db965f87474 | 7.5 | 0 |
"""EventStore 单元测试:追加写、轮转(P3 修复)、累计统计持久化(FR-5 修订)。"""
import os
import sys
import tempfile
import unittest
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from llm_sanitizer.events import EventStore, read_stats_file, tail_events
class TestEventStore(unittest.TestCase):
def test_... | JunyuZhan/llm-sanitizer | tests/test_events.py | .py | 77973a5af6c8177d | 7.5 | 0 |
"""格式处理器单元测试:docx/xlsx/pdf 保留格式脱敏/还原(v0.2+/v0.3)。
构造最小 docx/xlsx/pdf(ZIP+XML / FlateDecode stream),断言:文本被替换、
XML 结构与属性(如 xml:space)完好、PDF /Length 更新且还原后与原文一致。
"""
import os
import re
import sys
import tempfile
import unittest
import zipfile
import zlib
from pathlib import Path
sys.path.insert(0, os.path.dirname(os.p... | JunyuZhan/llm-sanitizer | tests/test_formats.py | .py | 66845b9fc8bfa368 | 7.5 | 0 |
"""masker 规则引擎单元测试:命中/误报/一致性/持久化/开关/并发。"""
import os
import sys
import tempfile
import threading
import unittest
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
from llm_sanitizer.masker import Masker, mask_text, restore_text
# 命中样例:(文本, 期望类别)
HIT_CASES = [
("身份证 1101011990010112... | JunyuZhan/llm-sanitizer | tests/test_masker.py | .py | 8f0daae76c338367 | 7.5 | 0 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.