Commit 78b41d73 authored by o's avatar o
Browse files

fix: keep imported reliability scores consistent

parent 660ba3c8

README.md

+8 −2
Changes for README.md: 8 added lines, 2 removed lines.
Original line number Diff line number Diff line
@@ -220,11 +220,17 @@ metric results. The endpoint stores the imported result, returns a new 
`analysis_id`, and makes the existing summary, PDF, and Excel endpoints

available without rerunning the analysis pipeline.



Every top-level `results[].score`, including Reliability, is a higher-is-better

quality score. With `score_scale: "ratio"`, the API converts `0..1` to

`0..100`. For Reliability, keep the raw CV percentage in

`detail.rel_overall_cv`; dashboard views use it only as the displayed CV basis

and do not grade the imported quality score a second time.



```json

{

  "source": "luc",

  "source": "ngsild-wrapper",

  "dataset_id": "stellio-air-quality-001",

  "report_name": "Luc data quality report",

  "report_name": "NGSI-LD data quality report",

  "score_scale": "ratio",

  "weighting_method": "custom",

  "metric_weights": {

api/schemas/analysis.py

+2 −0
Changes for api/schemas/analysis.py: 2 added lines, 0 removed lines.
Original line number Diff line number Diff line
@@ -30,6 +30,7 @@ class MetricSummary(BaseModel): 
    score: Optional[float] = None

    status: str                               # "evaluated", "n/a", "error"

    reason: Optional[str] = None              # N/A인 경우 사유

    score_semantics: Optional[str] = None     # e.g. quality_percent for imported scores

    summary: dict[str, Any] = {}              # 지표별 요약 데이터
@@ -51,6 +52,7 @@ class MetricDetailResponse(BaseModel): 
    analysis_id: str

    metric: str

    score: Optional[float] = None

    score_semantics: Optional[str] = None

    detail: dict[str, Any] = {}               # 지표별 상세 데이터

    rules: Optional[list[dict]] = None        # 규칙별 결과 (accuracy, consistency)

    samples: Optional[list[dict]] = None      # 샘플 데이터

api/services/analysis_service.py

+17 −0
Changes for api/services/analysis_service.py: 17 added lines, 0 removed lines.
Original line number Diff line number Diff line
@@ -277,6 +277,9 @@ class AnalysisService: 


            imported_results[metric] = {

                "score": score,

                # Imported top-level scores are normalized to 0..100 quality

                # scores. Reliability's raw CV remains in rel_overall_cv.

                "score_semantics": "quality_percent",

                "status": item.get("status", "evaluated"),

                "reason": item.get("reason"),

                "summary": item.get("summary", {}) or {},
@@ -900,11 +903,18 @@ class AnalysisService: 
        if not stored:

            return None



        imported_quality_scores = bool(

            (stored.get("config") or {}).get("__scores_are_quality_percent")

        )

        summaries = []

        for metric, data in stored["results"].items():

            summaries.append(MetricSummary(

                metric=metric,

                score=data.get("score"),

                score_semantics=(

                    data.get("score_semantics")

                    or ("quality_percent" if imported_quality_scores else None)

                ),

                status=data.get("status", "evaluated"),

                reason=data.get("reason"),

                summary=data.get("summary", {}),
@@ -935,6 +945,9 @@ class AnalysisService: 


        data = stored["results"][metric]

        detail = data.get("detail", {})

        imported_quality_scores = bool(

            (stored.get("config") or {}).get("__scores_are_quality_percent")

        )



        # 규칙별 결과 추출 (accuracy, consistency)

        rules = None
@@ -947,6 +960,10 @@ class AnalysisService: 
            analysis_id=analysis_id,

            metric=metric,

            score=data.get("score"),

            score_semantics=(

                data.get("score_semantics")

                or ("quality_percent" if imported_quality_scores else None)

            ),

            detail=detail,

            rules=rules,

        )

etsi-dq-frontend/package.json

+1 −0
Changes for etsi-dq-frontend/package.json: 1 added line, 0 removed lines.
Original line number Diff line number Diff line
@@ -6,6 +6,7 @@ 
  "scripts": {

    "dev": "vite",

    "build": "vite build",

    "test": "node tests/reliability-utils.test.mjs",

    "lint": "eslint .",

    "preview": "vite preview"

  },

etsi-dq-frontend/src/pages/AnalysisPage.jsx

+18 −16
Changes for etsi-dq-frontend/src/pages/AnalysisPage.jsx: 18 added lines, 16 removed lines.
Original line number Diff line number Diff line
@@ -3,6 +3,10 @@ import { useNavigate, useSearchParams } from "react-router-dom"; 
import Card from "../components/ui/Card";

import PageHeader from "../components/layout/PageHeader";

import { fetchAnalysisSummary } from "../api/client";

import {

  getReliabilityCvBasis,

  getReliabilityQualityScore,

} from "../utils/reliability";



const METRIC_LABELS = {

  completeness: "Completeness",
@@ -79,17 +83,6 @@ function getMetricScore(metricData) { 
  );

}



function getReliabilityQualityScore(cvValue) {

  const cv = Number(cvValue);

  if (!Number.isFinite(cv)) return 0;



  if (cv <= 5) return 100;

  if (cv <= 15) return 80;

  if (cv <= 30) return 60;

  if (cv <= 50) return 40;

  return 20;

}



function formatMetricValue(metric) {

  if (metric.rawStatus === "n/a") return "N/A";

  return formatScore(metric.score);
@@ -343,7 +336,7 @@ function getMetricDetailRows(metric, detailResponse = null) { 
      ["Evaluated Total", summary?.cons_checked_total ?? summary?.evaluated_total ?? "-"],

      ["Passed Total", summary?.cons_passed_total ?? summary?.passed_total ?? "-"],

      ["Violations", summary?.cons_violations_total ?? summary?.violations_total ?? "-"],

      ["Mode", summary?.cons_mode_used ?? (summary?.modes_used ?? []).join(", ") || "-"]

      ["Mode", summary?.cons_mode_used ?? ((summary?.modes_used ?? []).join(", ") || "-")]

    );

  }
@@ -380,7 +373,7 @@ function getMetricDetailRows(metric, detailResponse = null) { 
      ["Duplicate Rate", summary?.uniq_duplicate_rate !== undefined ? `${formatScore(Number(summary.uniq_duplicate_rate) * 100)}%` : summary?.duplicate_rate !== undefined ? `${formatScore(Number(summary.duplicate_rate) * 100)}%` : "-"],

      ["Mode", summary?.uniq_mode_label ?? "-"],

      ["Basis", summary?.uniq_basis_label ?? summary?.basis ?? "-"],

      ["Key Columns", summary?.uniq_key_caption ?? (summary?.keys ?? []).join(", ") || "-"]

      ["Key Columns", summary?.uniq_key_caption ?? ((summary?.keys ?? []).join(", ") || "-")]

    );

  }
@@ -471,8 +464,12 @@ export default function AnalysisPage() { 
      const rawScore = rawStatus === "n/a" ? 0 : getMetricScore(metricData);

      const score =

        metricKey === "reliability"

          ? getReliabilityQualityScore(rawScore)

          ? (getReliabilityQualityScore(metricData, rawScore) ?? 0)

          : rawScore;

      const cvBasis =

        metricKey === "reliability"

          ? getReliabilityCvBasis(metricData, rawScore)

          : null;

      const status = getMetricDisplayStatus(metricData, score, metricKey);

      const reason = getMetricReason(metricData);
@@ -481,6 +478,7 @@ export default function AnalysisPage() { 
        label: METRIC_LABELS[metricKey] ?? metricKey,

        score,

        rawScore,

        cvBasis,

        status,

        rawStatus,

        reason,
@@ -633,7 +631,9 @@ export default function AnalysisPage() { 
                )}

                {metric.key === "reliability" && metric.rawStatus !== "n/a" && (

                  <p className="mt-2 text-[10px] text-slate-400">

                    Calculated using CV basis: {formatScore(metric.rawScore)}%

                    {metric.cvBasis !== null

                      ? `Calculated using CV basis: ${formatScore(metric.cvBasis)}%`

                      : "CV basis unavailable"}

                  </p>

                )}

                {metric.reason && (
@@ -693,7 +693,9 @@ export default function AnalysisPage() { 
              <div className="rounded-xl bg-slate-50 p-3">

                <p className="text-xs font-semibold text-slate-500">Reliability Basis</p>

                <p className="mt-1 text-xl font-bold text-slate-900">

                  {formatScore(metricMap.reliability.rawScore)}%

                  {metricMap.reliability.cvBasis !== null

                    ? `${formatScore(metricMap.reliability.cvBasis)}%`

                    : "-"}

                </p>

                <p className="mt-1 text-xs text-slate-400">

                  CV basis{reliabilitySummary?.representative_column ? ` · ${reliabilitySummary.representative_column}` : ""}