Merge pull request #1 from ohquadri/patch-1

Created Directories

pyproject.toml

@@ -0,0 +1,11 @@
+[build-system]
+requires = ["setuptools>=68"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "acmecli"
+version = "0.0.1"
+requires-python = ">=3.10"
+
+[tool.setuptools.packages.find]
+where = ["src"]

run

@@ -0,0 +1,49 @@
+#!/usr/bin/env python3
+import sys, subprocess, re, os
+
+def do_install():
+    subprocess.check_call([sys.executable, "-m", "pip", "install", "-e", "."])
+    return 0
+
+def do_test():
+    import re, subprocess, sys, os, pathlib
+
+    # Always run from repo root
+    os.chdir(pathlib.Path(__file__).parent.resolve())
+
+    cmd = [sys.executable, "-m", "pytest", "tests",
+           "--maxfail=1", "--disable-warnings",
+           "--cov=acmecli", "--cov-report=term-missing"]
+    r = subprocess.run(cmd, text=True, capture_output=True)
+    out = (r.stdout or "") + (r.stderr or "")
+
+    # Extract counts
+    collected = re.search(r"collected\s+(\d+)", out)
+    passed    = re.search(r"(\d+)\s+passed", out)
+    cov       = re.search(r"TOTAL\s+.*?(\d+)%", out)
+
+    x = int(passed.group(1)) if passed else 0
+    y = int(collected.group(1)) if collected else 0
+    z = int(cov.group(1)) if cov else 0
+
+    print(f"{x}/{y} test cases passed. {z}% line coverage achieved.")
+    return 0 if r.returncode == 0 else 1
+
+
+def main():
+    if len(sys.argv) < 2:
+        print("Usage: run install|test|score <URL_FILE>")
+        sys.exit(1)
+    cmd = sys.argv[1]
+    if cmd == "install": sys.exit(do_install())
+    if cmd == "test":    sys.exit(do_test())
+    if cmd == "score":
+        if len(sys.argv) < 3: 
+            print("Usage: run score <URL_FILE>")
+            sys.exit(1)
+        sys.exit(do_score(sys.argv[2]))
+    print("Unknown command.")
+    sys.exit(1)
+
+if __name__ == "__main__":
+    main()

run.py

@@ -0,0 +1 @@
+import run

src/acmecli.egg-info/PKG-INFO

@@ -0,0 +1,4 @@
+Metadata-Version: 2.4
+Name: acmecli
+Version: 0.0.1
+Requires-Python: >=3.10

src/acmecli.egg-info/SOURCES.txt

@@ -0,0 +1,15 @@
+pyproject.toml
+src/acmecli/__init__.py
+src/acmecli/cli.py
+src/acmecli/reporter.py
+src/acmecli/scoring.py
+src/acmecli/types.py
+src/acmecli.egg-info/PKG-INFO
+src/acmecli.egg-info/SOURCES.txt
+src/acmecli.egg-info/dependency_links.txt
+src/acmecli.egg-info/top_level.txt
+src/acmecli/metrics/__init__.py
+src/acmecli/metrics/base.py
+src/acmecli/metrics/license_metric.py
+tests/test_metrics_contract.py
+tests/test_reporter_schema.py

src/acmecli.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

src/acmecli.egg-info/top_level.txt

@@ -0,0 +1 @@
+acmecli

src/acmecli/cli.py

@@ -0,0 +1,42 @@
+import sys
+from pathlib import Path
+from .types import ReportRow
+from .reporter import write_ndjson
+
+def _classify(url: str) -> str:
+    u = url.strip().lower()
+    if "huggingface.co/datasets/" in u:
+        return "DATASET"
+    if "github.com/" in u:
+        return "CODE"
+    if "huggingface.co/" in u:
+        return "MODEL"
+    return "CODE"
+
+def _stub_row(name: str) -> ReportRow:
+    zero = 0.0
+    return ReportRow(
+        name=name, category="MODEL",
+        net_score=zero, net_score_latency=0,
+        ramp_up_time=zero, ramp_up_time_latency=0,
+        bus_factor=zero, bus_factor_latency=0,
+        performance_claims=zero, performance_claims_latency=0,
+        license=zero, license_latency=0,
+        size_score={"raspberry_pi":0.0, "jetson_nano":0.0, "desktop_pc":0.0, "aws_server":0.0},
+        size_score_latency=0,
+        dataset_and_code_score=zero, dataset_and_code_score_latency=0,
+        dataset_quality=zero, dataset_quality_latency=0,
+        code_quality=zero, code_quality_latency=0
+    )
+
+def main(argv: list[str]) -> int:
+    # argv pattern: ["score", "/abs/path/URL_FILE"]
+    _, url_file = argv
+    lines = Path(url_file).read_text(encoding="utf-8").splitlines()
+    for raw in lines:
+        url = raw.strip()
+        if not url:
+            continue
+        if _classify(url) == "MODEL":
+            write_ndjson(_stub_row(url))
+    return 0

src/acmecli/metrics/base.py

@@ -0,0 +1,7 @@
+from typing import List
+from ..types import Metric, MetricValue, TargetSpec, SourceHandler, Cache
+
+REGISTRY: List[Metric] = []
+
+def register(metric: Metric) -> None:
+    REGISTRY.append(metric)

src/acmecli/metrics/license_metric.py

@@ -0,0 +1,18 @@
+import time
+from ..types import Metric, Signals, TargetSpec, SourceHandler, Cache, MetricValue
+from .base import register
+
+class LicenseMetric:
+    name = "license"
+
+    def collect(self, spec: TargetSpec, handler: SourceHandler, cache: Cache) -> Signals:
+        # Week 1 stub: no network; just return empty signals
+        return {}
+
+    def score(self, signals: Signals) -> MetricValue:
+        t0 = time.perf_counter()
+        value = 0.0  # TODO: map license presence/compatibility → [0,1]
+        latency_ms = int((time.perf_counter() - t0) * 1000)
+        return MetricValue(self.name, value, latency_ms)
+
+register(LicenseMetric())

src/acmecli/reporter.py

@@ -0,0 +1,6 @@
+import json
+from dataclasses import asdict
+from .types import ReportRow
+
+def write_ndjson(row: ReportRow) -> None:
+    print(json.dumps(asdict(row), ensure_ascii=False))

src/acmecli/scoring.py

@@ -0,0 +1,26 @@
+from typing import Dict, List
+from .types import MetricValue
+
+# Initial weights (sum ≈ 1.0). Tweak later.
+WEIGHTS: Dict[str, float] = {
+    "ramp_up_time": 0.15,
+    "bus_factor": 0.15,
+    "performance_claims": 0.10,
+    "license": 0.20,
+    "size": 0.10,  # average of size_score dict
+    "dataset_and_code_score": 0.10,
+    "dataset_quality": 0.10,
+    "code_quality": 0.10,
+}
+
+class ScoringEngine:
+    def compute(self, values: List[MetricValue], size_avg: float = 0.0) -> tuple[float, int]:
+        lookup = {mv.name: mv for mv in values}
+        total_latency = sum(mv.latency_ms for mv in values)
+        net = 0.0
+        for k, w in WEIGHTS.items():
+            if k == "size":
+                net += w * float(size_avg)
+            elif k in lookup:
+                net += w * float(lookup[k].value)
+        return (max(0.0, min(1.0, net)), total_latency)

src/acmecli/types.py

@@ -0,0 +1,64 @@
+from dataclasses import dataclass, asdict
+from typing import Protocol, Iterable, TypedDict, Literal, Optional, Dict
+
+Category = Literal["MODEL", "DATASET", "CODE"]
+Source   = Literal["GITHUB", "HUGGINGFACE", "LOCAL"]
+
+@dataclass(frozen=True)
+class TargetSpec:
+    url: str
+    source: Source
+    name: str
+    category: Category
+    revision: Optional[str] = None  # commit/tag if known
+
+class Signals(TypedDict, total=False):
+    readme_text: str
+    license_name: str
+    contributors: Dict[str, int]
+    stars: int
+    downloads: int
+
+@dataclass(frozen=True)
+class MetricValue:
+    name: str              # e.g. "ramp_up_time"
+    value: float           # [0,1]
+    latency_ms: int
+
+@dataclass(frozen=True)
+class ReportRow:
+    # Required NDJSON fields + per-metric latencies (values are stubs for now)
+    name: str
+    category: Category
+    net_score: float
+    net_score_latency: int
+    ramp_up_time: float
+    ramp_up_time_latency: int
+    bus_factor: float
+    bus_factor_latency: int
+    performance_claims: float
+    performance_claims_latency: int
+    license: float
+    license_latency: int
+    size_score: Dict[str, float]     # {raspberry_pi, jetson_nano, desktop_pc, aws_server}
+    size_score_latency: int
+    dataset_and_code_score: float
+    dataset_and_code_score_latency: int
+    dataset_quality: float
+    dataset_quality_latency: int
+    code_quality: float
+    code_quality_latency: int
+
+class SourceHandler(Protocol):
+    def resolve_revision(self, url: str) -> str: ...
+    def fetch_meta(self, spec: TargetSpec) -> dict: ...
+    def stream_files(self, spec: TargetSpec, patterns: list[str]) -> Iterable[tuple[str, bytes]]: ...
+
+class Cache(Protocol):
+    def get(self, key: str) -> bytes | None: ...
+    def set(self, key: str, data: bytes, etag: str | None = None) -> None: ...
+
+class Metric(Protocol):
+    name: str
+    def collect(self, spec: TargetSpec, handler: SourceHandler, cache: Cache) -> Signals: ...
+    def score(self, signals: Signals) -> MetricValue: ...

tests/test_metrics_contract.py

@@ -0,0 +1,11 @@
+from acmecli.metrics.base import REGISTRY
+from acmecli.metrics.license_metric import LicenseMetric
+
+def test_registry_has_license_metric():
+    assert any(m.name == "license" for m in REGISTRY)
+
+def test_metric_value_range():
+    m = LicenseMetric()
+    mv = m.score({})
+    assert 0.0 <= mv.value <= 1.0
+    assert mv.latency_ms >= 0

tests/test_reporter_schema.py

@@ -0,0 +1,16 @@
+from acmecli.types import ReportRow
+
+def test_reportrow_has_required_fields():
+    row = ReportRow(
+        name="demo", category="MODEL",
+        net_score=0.0, net_score_latency=0,
+        ramp_up_time=0.0, ramp_up_time_latency=0,
+        bus_factor=0.0, bus_factor_latency=0,
+        performance_claims=0.0, performance_claims_latency=0,
+        license=0.0, license_latency=0,
+        size_score={}, size_score_latency=0,
+        dataset_and_code_score=0.0, dataset_and_code_score_latency=0,
+        dataset_quality=0.0, dataset_quality_latency=0,
+        code_quality=0.0, code_quality_latency=0
+    )
+    assert row.category == "MODEL"