@contractspec/lib.provider-ranking 0.7.5 → 0.7.8

package/README.md

@@ -1,44 +1,77 @@
 # @contractspec/lib.provider-ranking
 
-Website: https://contractspec.io/
+Website: https://contractspec.io
 
 **AI provider ranking: benchmark ingestion, scoring, and model comparison.**
 
-Ingests benchmark data from multiple sources (Chatbot Arena, SWE-bench, Artificial Analysis, Open LLM Leaderboard), normalizes scores to a 0-100 scale, and computes composite rankings across dimensions like coding, reasoning, cost, and latency.
+## What It Provides
+
+- **Layer**: lib.
+- **Consumers**: module.provider-ranking.
+- Related ContractSpec packages include `@contractspec/tool.bun`, `@contractspec/tool.typescript`.
+- Related ContractSpec packages include `@contractspec/tool.bun`, `@contractspec/tool.typescript`.
 
 ## Installation
 
-```bash
-bun add @contractspec/lib.provider-ranking
-```
+`npm install @contractspec/lib.provider-ranking`
 
-## Exports
+or
 
-- `.` -- Core types, store interface, and in-memory store
-- `./types` -- `BenchmarkResult`, `ModelRanking`, `ModelProfile`, `BenchmarkDimension`, `DimensionWeightConfig`
-- `./store` -- `ProviderRankingStore` interface
-- `./in-memory-store` -- `InMemoryProviderRankingStore` class
-- `./scoring` -- `computeModelRankings()`, `normalizeScore()`, `DEFAULT_DIMENSION_WEIGHTS`
-- `./ingesters` -- `chatbotArenaIngester`, `sweBenchIngester`, `artificialAnalysisIngester`, `IngesterRegistry`
-- `./eval` -- `EvalRunner`, `EvalSuite`, `EvalCase` for custom evaluation
+`bun add @contractspec/lib.provider-ranking`
 
 ## Usage
 
-```ts
-import { InMemoryProviderRankingStore } from "@contractspec/lib.provider-ranking/in-memory-store";
-import { createDefaultIngesterRegistry } from "@contractspec/lib.provider-ranking/ingesters";
-import { computeModelRankings } from "@contractspec/lib.provider-ranking/scoring";
+Import the root entrypoint from `@contractspec/lib.provider-ranking`, or choose a documented subpath when you only need one part of the package surface.
+
+## Architecture
+
+- `src/eval` is part of the package's public or composition surface.
+- `src/in-memory-store.ts` is part of the package's public or composition surface.
+- `src/index.ts` is the root public barrel and package entrypoint.
+- `src/ingesters` is part of the package's public or composition surface.
+- `src/scoring` is part of the package's public or composition surface.
+- `src/store.ts` is part of the package's public or composition surface.
+- `src/types.ts` is shared public type definitions.
+
+## Public Entry Points
+
+- Export `.` resolves through `./src/index.ts`.
+- Export `./eval` resolves through `./src/eval/index.ts`.
+- Export `./eval/runner` resolves through `./src/eval/runner.ts`.
+- Export `./eval/types` resolves through `./src/eval/types.ts`.
+- Export `./in-memory-store` resolves through `./src/in-memory-store.ts`.
+- Export `./ingesters` resolves through `./src/ingesters/index.ts`.
+- Export `./ingesters/artificial-analysis` resolves through `./src/ingesters/artificial-analysis.ts`.
+- Export `./ingesters/chatbot-arena` resolves through `./src/ingesters/chatbot-arena.ts`.
+- Export `./ingesters/fetch-utils` resolves through `./src/ingesters/fetch-utils.ts`.
+- Export `./ingesters/open-llm-leaderboard` resolves through `./src/ingesters/open-llm-leaderboard.ts`.
+- The package publishes 19 total export subpaths; keep docs aligned with `package.json`.
+
+## Local Commands
+
+- `bun run dev` — contractspec-bun-build dev
+- `bun run build` — bun run prebuild && bun run build:bundle && bun run build:types
+- `bun run test` — bun test --pass-with-no-tests
+- `bun run lint` — bun lint:fix
+- `bun run lint:check` — biome check .
+- `bun run lint:fix` — biome check --write --unsafe --only=nursery/useSortedClasses . && biome check --write .
+- `bun run typecheck` — tsc --noEmit
+- `bun run publish:pkg` — bun publish --tolerate-republish --ignore-scripts --verbose
+- `bun run publish:pkg:canary` — bun publish:pkg --tag canary
+- `bun run clean` — rimraf dist .turbo
+- `bun run build:bundle` — contractspec-bun-build transpile
+- `bun run build:types` — contractspec-bun-build types
+- `bun run prebuild` — contractspec-bun-build prebuild
 
-const store = new InMemoryProviderRankingStore();
-const registry = createDefaultIngesterRegistry();
+## Recent Updates
 
-const ingester = registry.get("swe-bench");
-const results = await ingester.ingest();
+- Replace eslint+prettier by biomejs to optimize speed.
+- Resolve lint, build, and type errors across nine packages.
+- Add first-class transport, auth, versioning, and BYOK support across all integrations.
+- Add AI provider ranking system with ranking-driven model selection.
 
-for (const result of results) {
-  await store.addBenchmarkResult(result);
-}
+## Notes
 
-const rankings = computeModelRankings(await store.listBenchmarkResults({}));
-console.log(rankings);
-```
+- Store interface is the adapter boundary — do not leak implementation details.
+- Scoring algorithms must stay deterministic (no randomness, no side effects).
+- Benchmark dimension enum is shared across ingesters and scoring — keep in sync.

package/dist/browser/ingesters/index.js

@@ -249,73 +249,6 @@ function mapOrganizationToProvider(org) {
   return org;
 }
 
-// src/ingesters/swe-bench.ts
-var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
-var sweBenchIngester = {
-  source: "swe-bench",
-  displayName: "SWE-bench",
-  description: "Software engineering task completion rates from SWE-bench.",
-  async ingest(options) {
-    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
-      return [];
-    }
-    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
-    const response = await fetchWithRetry(url, { fetch: options?.fetch });
-    const text = await response.text();
-    const data = parseJsonSafe(text, "SWE-bench");
-    const now = new Date;
-    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
-    if (options?.modelFilter?.length) {
-      const filterSet = new Set(options.modelFilter);
-      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
-    }
-    if (options?.maxResults) {
-      entries = entries.slice(0, options.maxResults);
-    }
-    let results = entries.map((entry) => {
-      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
-      const org = entry.organization?.toLowerCase() ?? "unknown";
-      return {
-        id: `swe-bench:${modelId}:coding`,
-        modelId,
-        providerKey: mapOrganizationToProvider2(org),
-        source: "swe-bench",
-        dimension: "coding",
-        score: Math.max(0, Math.min(100, entry.resolved_rate)),
-        rawScore: entry.resolved_rate,
-        metadata: {
-          organization: entry.organization,
-          date: entry.date
-        },
-        measuredAt: entry.date ? new Date(entry.date) : now,
-        ingestedAt: now
-      };
-    });
-    const { fromDate, toDate } = options ?? {};
-    if (fromDate) {
-      results = results.filter((r) => r.measuredAt >= fromDate);
-    }
-    if (toDate) {
-      results = results.filter((r) => r.measuredAt <= toDate);
-    }
-    return results;
-  }
-};
-function mapOrganizationToProvider2(org) {
-  const normalized = org.toLowerCase();
-  if (normalized.includes("openai"))
-    return "openai";
-  if (normalized.includes("anthropic"))
-    return "anthropic";
-  if (normalized.includes("google") || normalized.includes("deepmind"))
-    return "gemini";
-  if (normalized.includes("mistral"))
-    return "mistral";
-  if (normalized.includes("meta"))
-    return "meta";
-  return org;
-}
-
 // src/ingesters/open-llm-leaderboard.ts
 var DEFAULT_HF_URL = "https://huggingface.co/api/spaces/open-llm-leaderboard/open_llm_leaderboard/results";
 var BENCHMARK_MAPPINGS = [
@@ -344,7 +277,7 @@ var openLlmLeaderboardIngester = {
     for (const entry of entries) {
       const modelId = entry.model_name.toLowerCase().replace(/\s+/g, "-");
       const org = entry.organization?.toLowerCase() ?? "unknown";
-      const providerKey = mapOrganizationToProvider3(org);
+      const providerKey = mapOrganizationToProvider2(org);
       for (const mapping of BENCHMARK_MAPPINGS) {
         if (dims && !dims.has(mapping.dimension))
           continue;
@@ -371,6 +304,73 @@ var openLlmLeaderboardIngester = {
     return options?.maxResults ? results.slice(0, options.maxResults) : results;
   }
 };
+function mapOrganizationToProvider2(org) {
+  const normalized = org.toLowerCase();
+  if (normalized.includes("openai"))
+    return "openai";
+  if (normalized.includes("anthropic"))
+    return "anthropic";
+  if (normalized.includes("google") || normalized.includes("deepmind"))
+    return "gemini";
+  if (normalized.includes("mistral"))
+    return "mistral";
+  if (normalized.includes("meta"))
+    return "meta";
+  return org;
+}
+
+// src/ingesters/swe-bench.ts
+var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
+var sweBenchIngester = {
+  source: "swe-bench",
+  displayName: "SWE-bench",
+  description: "Software engineering task completion rates from SWE-bench.",
+  async ingest(options) {
+    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
+      return [];
+    }
+    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
+    const response = await fetchWithRetry(url, { fetch: options?.fetch });
+    const text = await response.text();
+    const data = parseJsonSafe(text, "SWE-bench");
+    const now = new Date;
+    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
+    if (options?.modelFilter?.length) {
+      const filterSet = new Set(options.modelFilter);
+      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
+    }
+    if (options?.maxResults) {
+      entries = entries.slice(0, options.maxResults);
+    }
+    let results = entries.map((entry) => {
+      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
+      const org = entry.organization?.toLowerCase() ?? "unknown";
+      return {
+        id: `swe-bench:${modelId}:coding`,
+        modelId,
+        providerKey: mapOrganizationToProvider3(org),
+        source: "swe-bench",
+        dimension: "coding",
+        score: Math.max(0, Math.min(100, entry.resolved_rate)),
+        rawScore: entry.resolved_rate,
+        metadata: {
+          organization: entry.organization,
+          date: entry.date
+        },
+        measuredAt: entry.date ? new Date(entry.date) : now,
+        ingestedAt: now
+      };
+    });
+    const { fromDate, toDate } = options ?? {};
+    if (fromDate) {
+      results = results.filter((r) => r.measuredAt >= fromDate);
+    }
+    if (toDate) {
+      results = results.filter((r) => r.measuredAt <= toDate);
+    }
+    return results;
+  }
+};
 function mapOrganizationToProvider3(org) {
   const normalized = org.toLowerCase();
   if (normalized.includes("openai"))

package/dist/browser/ingesters/registry.js

@@ -249,73 +249,6 @@ function mapOrganizationToProvider(org) {
   return org;
 }
 
-// src/ingesters/swe-bench.ts
-var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
-var sweBenchIngester = {
-  source: "swe-bench",
-  displayName: "SWE-bench",
-  description: "Software engineering task completion rates from SWE-bench.",
-  async ingest(options) {
-    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
-      return [];
-    }
-    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
-    const response = await fetchWithRetry(url, { fetch: options?.fetch });
-    const text = await response.text();
-    const data = parseJsonSafe(text, "SWE-bench");
-    const now = new Date;
-    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
-    if (options?.modelFilter?.length) {
-      const filterSet = new Set(options.modelFilter);
-      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
-    }
-    if (options?.maxResults) {
-      entries = entries.slice(0, options.maxResults);
-    }
-    let results = entries.map((entry) => {
-      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
-      const org = entry.organization?.toLowerCase() ?? "unknown";
-      return {
-        id: `swe-bench:${modelId}:coding`,
-        modelId,
-        providerKey: mapOrganizationToProvider2(org),
-        source: "swe-bench",
-        dimension: "coding",
-        score: Math.max(0, Math.min(100, entry.resolved_rate)),
-        rawScore: entry.resolved_rate,
-        metadata: {
-          organization: entry.organization,
-          date: entry.date
-        },
-        measuredAt: entry.date ? new Date(entry.date) : now,
-        ingestedAt: now
-      };
-    });
-    const { fromDate, toDate } = options ?? {};
-    if (fromDate) {
-      results = results.filter((r) => r.measuredAt >= fromDate);
-    }
-    if (toDate) {
-      results = results.filter((r) => r.measuredAt <= toDate);
-    }
-    return results;
-  }
-};
-function mapOrganizationToProvider2(org) {
-  const normalized = org.toLowerCase();
-  if (normalized.includes("openai"))
-    return "openai";
-  if (normalized.includes("anthropic"))
-    return "anthropic";
-  if (normalized.includes("google") || normalized.includes("deepmind"))
-    return "gemini";
-  if (normalized.includes("mistral"))
-    return "mistral";
-  if (normalized.includes("meta"))
-    return "meta";
-  return org;
-}
-
 // src/ingesters/open-llm-leaderboard.ts
 var DEFAULT_HF_URL = "https://huggingface.co/api/spaces/open-llm-leaderboard/open_llm_leaderboard/results";
 var BENCHMARK_MAPPINGS = [
@@ -344,7 +277,7 @@ var openLlmLeaderboardIngester = {
     for (const entry of entries) {
       const modelId = entry.model_name.toLowerCase().replace(/\s+/g, "-");
       const org = entry.organization?.toLowerCase() ?? "unknown";
-      const providerKey = mapOrganizationToProvider3(org);
+      const providerKey = mapOrganizationToProvider2(org);
       for (const mapping of BENCHMARK_MAPPINGS) {
         if (dims && !dims.has(mapping.dimension))
           continue;
@@ -371,6 +304,73 @@ var openLlmLeaderboardIngester = {
     return options?.maxResults ? results.slice(0, options.maxResults) : results;
   }
 };
+function mapOrganizationToProvider2(org) {
+  const normalized = org.toLowerCase();
+  if (normalized.includes("openai"))
+    return "openai";
+  if (normalized.includes("anthropic"))
+    return "anthropic";
+  if (normalized.includes("google") || normalized.includes("deepmind"))
+    return "gemini";
+  if (normalized.includes("mistral"))
+    return "mistral";
+  if (normalized.includes("meta"))
+    return "meta";
+  return org;
+}
+
+// src/ingesters/swe-bench.ts
+var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
+var sweBenchIngester = {
+  source: "swe-bench",
+  displayName: "SWE-bench",
+  description: "Software engineering task completion rates from SWE-bench.",
+  async ingest(options) {
+    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
+      return [];
+    }
+    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
+    const response = await fetchWithRetry(url, { fetch: options?.fetch });
+    const text = await response.text();
+    const data = parseJsonSafe(text, "SWE-bench");
+    const now = new Date;
+    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
+    if (options?.modelFilter?.length) {
+      const filterSet = new Set(options.modelFilter);
+      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
+    }
+    if (options?.maxResults) {
+      entries = entries.slice(0, options.maxResults);
+    }
+    let results = entries.map((entry) => {
+      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
+      const org = entry.organization?.toLowerCase() ?? "unknown";
+      return {
+        id: `swe-bench:${modelId}:coding`,
+        modelId,
+        providerKey: mapOrganizationToProvider3(org),
+        source: "swe-bench",
+        dimension: "coding",
+        score: Math.max(0, Math.min(100, entry.resolved_rate)),
+        rawScore: entry.resolved_rate,
+        metadata: {
+          organization: entry.organization,
+          date: entry.date
+        },
+        measuredAt: entry.date ? new Date(entry.date) : now,
+        ingestedAt: now
+      };
+    });
+    const { fromDate, toDate } = options ?? {};
+    if (fromDate) {
+      results = results.filter((r) => r.measuredAt >= fromDate);
+    }
+    if (toDate) {
+      results = results.filter((r) => r.measuredAt <= toDate);
+    }
+    return results;
+  }
+};
 function mapOrganizationToProvider3(org) {
   const normalized = org.toLowerCase();
   if (normalized.includes("openai"))

package/dist/eval/index.d.ts

@@ -1,2 +1,2 @@
-export type { EvalCase, EvalSuite, EvalCaseResult, EvalRunResult, EvalLLMAdapter, } from './types';
 export { EvalRunner } from './runner';
+export type { EvalCase, EvalCaseResult, EvalLLMAdapter, EvalRunResult, EvalSuite, } from './types';

package/dist/index.d.ts

@@ -1,4 +1,4 @@
-export type { BenchmarkDimension, BenchmarkSource, BenchmarkResult, DimensionScore, ModelRanking, ModelProfile, BenchmarkResultQuery, BenchmarkResultListResult, RankingQuery, RankingListResult, IngestionRun, DimensionWeightConfig, } from './types';
-export { BENCHMARK_DIMENSIONS } from './types';
-export type { ProviderRankingStore } from './store';
 export { InMemoryProviderRankingStore } from './in-memory-store';
+export type { ProviderRankingStore } from './store';
+export type { BenchmarkDimension, BenchmarkResult, BenchmarkResultListResult, BenchmarkResultQuery, BenchmarkSource, DimensionScore, DimensionWeightConfig, IngestionRun, ModelProfile, ModelRanking, RankingListResult, RankingQuery, } from './types';
+export { BENCHMARK_DIMENSIONS } from './types';

package/dist/ingesters/index.d.ts

@@ -1,7 +1,7 @@
-export type { BenchmarkIngester, IngesterOptions } from './types';
-export { chatbotArenaIngester } from './chatbot-arena';
 export { artificialAnalysisIngester } from './artificial-analysis';
-export { sweBenchIngester } from './swe-bench';
-export { openLlmLeaderboardIngester } from './open-llm-leaderboard';
-export { IngesterRegistry, createDefaultIngesterRegistry } from './registry';
+export { chatbotArenaIngester } from './chatbot-arena';
 export { fetchWithRetry, parseJsonSafe } from './fetch-utils';
+export { openLlmLeaderboardIngester } from './open-llm-leaderboard';
+export { createDefaultIngesterRegistry, IngesterRegistry } from './registry';
+export { sweBenchIngester } from './swe-bench';
+export type { BenchmarkIngester, IngesterOptions } from './types';

package/dist/ingesters/index.js

@@ -250,73 +250,6 @@ function mapOrganizationToProvider(org) {
   return org;
 }
 
-// src/ingesters/swe-bench.ts
-var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
-var sweBenchIngester = {
-  source: "swe-bench",
-  displayName: "SWE-bench",
-  description: "Software engineering task completion rates from SWE-bench.",
-  async ingest(options) {
-    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
-      return [];
-    }
-    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
-    const response = await fetchWithRetry(url, { fetch: options?.fetch });
-    const text = await response.text();
-    const data = parseJsonSafe(text, "SWE-bench");
-    const now = new Date;
-    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
-    if (options?.modelFilter?.length) {
-      const filterSet = new Set(options.modelFilter);
-      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
-    }
-    if (options?.maxResults) {
-      entries = entries.slice(0, options.maxResults);
-    }
-    let results = entries.map((entry) => {
-      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
-      const org = entry.organization?.toLowerCase() ?? "unknown";
-      return {
-        id: `swe-bench:${modelId}:coding`,
-        modelId,
-        providerKey: mapOrganizationToProvider2(org),
-        source: "swe-bench",
-        dimension: "coding",
-        score: Math.max(0, Math.min(100, entry.resolved_rate)),
-        rawScore: entry.resolved_rate,
-        metadata: {
-          organization: entry.organization,
-          date: entry.date
-        },
-        measuredAt: entry.date ? new Date(entry.date) : now,
-        ingestedAt: now
-      };
-    });
-    const { fromDate, toDate } = options ?? {};
-    if (fromDate) {
-      results = results.filter((r) => r.measuredAt >= fromDate);
-    }
-    if (toDate) {
-      results = results.filter((r) => r.measuredAt <= toDate);
-    }
-    return results;
-  }
-};
-function mapOrganizationToProvider2(org) {
-  const normalized = org.toLowerCase();
-  if (normalized.includes("openai"))
-    return "openai";
-  if (normalized.includes("anthropic"))
-    return "anthropic";
-  if (normalized.includes("google") || normalized.includes("deepmind"))
-    return "gemini";
-  if (normalized.includes("mistral"))
-    return "mistral";
-  if (normalized.includes("meta"))
-    return "meta";
-  return org;
-}
-
 // src/ingesters/open-llm-leaderboard.ts
 var DEFAULT_HF_URL = "https://huggingface.co/api/spaces/open-llm-leaderboard/open_llm_leaderboard/results";
 var BENCHMARK_MAPPINGS = [
@@ -345,7 +278,7 @@ var openLlmLeaderboardIngester = {
     for (const entry of entries) {
       const modelId = entry.model_name.toLowerCase().replace(/\s+/g, "-");
       const org = entry.organization?.toLowerCase() ?? "unknown";
-      const providerKey = mapOrganizationToProvider3(org);
+      const providerKey = mapOrganizationToProvider2(org);
       for (const mapping of BENCHMARK_MAPPINGS) {
         if (dims && !dims.has(mapping.dimension))
           continue;
@@ -372,6 +305,73 @@ var openLlmLeaderboardIngester = {
     return options?.maxResults ? results.slice(0, options.maxResults) : results;
   }
 };
+function mapOrganizationToProvider2(org) {
+  const normalized = org.toLowerCase();
+  if (normalized.includes("openai"))
+    return "openai";
+  if (normalized.includes("anthropic"))
+    return "anthropic";
+  if (normalized.includes("google") || normalized.includes("deepmind"))
+    return "gemini";
+  if (normalized.includes("mistral"))
+    return "mistral";
+  if (normalized.includes("meta"))
+    return "meta";
+  return org;
+}
+
+// src/ingesters/swe-bench.ts
+var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
+var sweBenchIngester = {
+  source: "swe-bench",
+  displayName: "SWE-bench",
+  description: "Software engineering task completion rates from SWE-bench.",
+  async ingest(options) {
+    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
+      return [];
+    }
+    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
+    const response = await fetchWithRetry(url, { fetch: options?.fetch });
+    const text = await response.text();
+    const data = parseJsonSafe(text, "SWE-bench");
+    const now = new Date;
+    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
+    if (options?.modelFilter?.length) {
+      const filterSet = new Set(options.modelFilter);
+      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
+    }
+    if (options?.maxResults) {
+      entries = entries.slice(0, options.maxResults);
+    }
+    let results = entries.map((entry) => {
+      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
+      const org = entry.organization?.toLowerCase() ?? "unknown";
+      return {
+        id: `swe-bench:${modelId}:coding`,
+        modelId,
+        providerKey: mapOrganizationToProvider3(org),
+        source: "swe-bench",
+        dimension: "coding",
+        score: Math.max(0, Math.min(100, entry.resolved_rate)),
+        rawScore: entry.resolved_rate,
+        metadata: {
+          organization: entry.organization,
+          date: entry.date
+        },
+        measuredAt: entry.date ? new Date(entry.date) : now,
+        ingestedAt: now
+      };
+    });
+    const { fromDate, toDate } = options ?? {};
+    if (fromDate) {
+      results = results.filter((r) => r.measuredAt >= fromDate);
+    }
+    if (toDate) {
+      results = results.filter((r) => r.measuredAt <= toDate);
+    }
+    return results;
+  }
+};
 function mapOrganizationToProvider3(org) {
   const normalized = org.toLowerCase();
   if (normalized.includes("openai"))

package/dist/ingesters/registry.js

@@ -250,73 +250,6 @@ function mapOrganizationToProvider(org) {
   return org;
 }
 
-// src/ingesters/swe-bench.ts
-var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
-var sweBenchIngester = {
-  source: "swe-bench",
-  displayName: "SWE-bench",
-  description: "Software engineering task completion rates from SWE-bench.",
-  async ingest(options) {
-    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
-      return [];
-    }
-    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
-    const response = await fetchWithRetry(url, { fetch: options?.fetch });
-    const text = await response.text();
-    const data = parseJsonSafe(text, "SWE-bench");
-    const now = new Date;
-    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
-    if (options?.modelFilter?.length) {
-      const filterSet = new Set(options.modelFilter);
-      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
-    }
-    if (options?.maxResults) {
-      entries = entries.slice(0, options.maxResults);
-    }
-    let results = entries.map((entry) => {
-      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
-      const org = entry.organization?.toLowerCase() ?? "unknown";
-      return {
-        id: `swe-bench:${modelId}:coding`,
-        modelId,
-        providerKey: mapOrganizationToProvider2(org),
-        source: "swe-bench",
-        dimension: "coding",
-        score: Math.max(0, Math.min(100, entry.resolved_rate)),
-        rawScore: entry.resolved_rate,
-        metadata: {
-          organization: entry.organization,
-          date: entry.date
-        },
-        measuredAt: entry.date ? new Date(entry.date) : now,
-        ingestedAt: now
-      };
-    });
-    const { fromDate, toDate } = options ?? {};
-    if (fromDate) {
-      results = results.filter((r) => r.measuredAt >= fromDate);
-    }
-    if (toDate) {
-      results = results.filter((r) => r.measuredAt <= toDate);
-    }
-    return results;
-  }
-};
-function mapOrganizationToProvider2(org) {
-  const normalized = org.toLowerCase();
-  if (normalized.includes("openai"))
-    return "openai";
-  if (normalized.includes("anthropic"))
-    return "anthropic";
-  if (normalized.includes("google") || normalized.includes("deepmind"))
-    return "gemini";
-  if (normalized.includes("mistral"))
-    return "mistral";
-  if (normalized.includes("meta"))
-    return "meta";
-  return org;
-}
-
 // src/ingesters/open-llm-leaderboard.ts
 var DEFAULT_HF_URL = "https://huggingface.co/api/spaces/open-llm-leaderboard/open_llm_leaderboard/results";
 var BENCHMARK_MAPPINGS = [
@@ -345,7 +278,7 @@ var openLlmLeaderboardIngester = {
     for (const entry of entries) {
       const modelId = entry.model_name.toLowerCase().replace(/\s+/g, "-");
       const org = entry.organization?.toLowerCase() ?? "unknown";
-      const providerKey = mapOrganizationToProvider3(org);
+      const providerKey = mapOrganizationToProvider2(org);
       for (const mapping of BENCHMARK_MAPPINGS) {
         if (dims && !dims.has(mapping.dimension))
           continue;
@@ -372,6 +305,73 @@ var openLlmLeaderboardIngester = {
     return options?.maxResults ? results.slice(0, options.maxResults) : results;
   }
 };
+function mapOrganizationToProvider2(org) {
+  const normalized = org.toLowerCase();
+  if (normalized.includes("openai"))
+    return "openai";
+  if (normalized.includes("anthropic"))
+    return "anthropic";
+  if (normalized.includes("google") || normalized.includes("deepmind"))
+    return "gemini";
+  if (normalized.includes("mistral"))
+    return "mistral";
+  if (normalized.includes("meta"))
+    return "meta";
+  return org;
+}
+
+// src/ingesters/swe-bench.ts
+var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
+var sweBenchIngester = {
+  source: "swe-bench",
+  displayName: "SWE-bench",
+  description: "Software engineering task completion rates from SWE-bench.",
+  async ingest(options) {
+    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
+      return [];
+    }
+    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
+    const response = await fetchWithRetry(url, { fetch: options?.fetch });
+    const text = await response.text();
+    const data = parseJsonSafe(text, "SWE-bench");
+    const now = new Date;
+    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
+    if (options?.modelFilter?.length) {
+      const filterSet = new Set(options.modelFilter);
+      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
+    }
+    if (options?.maxResults) {
+      entries = entries.slice(0, options.maxResults);
+    }
+    let results = entries.map((entry) => {
+      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
+      const org = entry.organization?.toLowerCase() ?? "unknown";
+      return {
+        id: `swe-bench:${modelId}:coding`,
+        modelId,
+        providerKey: mapOrganizationToProvider3(org),
+        source: "swe-bench",
+        dimension: "coding",
+        score: Math.max(0, Math.min(100, entry.resolved_rate)),
+        rawScore: entry.resolved_rate,
+        metadata: {
+          organization: entry.organization,
+          date: entry.date
+        },
+        measuredAt: entry.date ? new Date(entry.date) : now,
+        ingestedAt: now
+      };
+    });
+    const { fromDate, toDate } = options ?? {};
+    if (fromDate) {
+      results = results.filter((r) => r.measuredAt >= fromDate);
+    }
+    if (toDate) {
+      results = results.filter((r) => r.measuredAt <= toDate);
+    }
+    return results;
+  }
+};
 function mapOrganizationToProvider3(org) {
   const normalized = org.toLowerCase();
   if (normalized.includes("openai"))

package/dist/node/ingesters/index.js

@@ -249,73 +249,6 @@ function mapOrganizationToProvider(org) {
   return org;
 }
 
-// src/ingesters/swe-bench.ts
-var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
-var sweBenchIngester = {
-  source: "swe-bench",
-  displayName: "SWE-bench",
-  description: "Software engineering task completion rates from SWE-bench.",
-  async ingest(options) {
-    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
-      return [];
-    }
-    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
-    const response = await fetchWithRetry(url, { fetch: options?.fetch });
-    const text = await response.text();
-    const data = parseJsonSafe(text, "SWE-bench");
-    const now = new Date;
-    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
-    if (options?.modelFilter?.length) {
-      const filterSet = new Set(options.modelFilter);
-      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
-    }
-    if (options?.maxResults) {
-      entries = entries.slice(0, options.maxResults);
-    }
-    let results = entries.map((entry) => {
-      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
-      const org = entry.organization?.toLowerCase() ?? "unknown";
-      return {
-        id: `swe-bench:${modelId}:coding`,
-        modelId,
-        providerKey: mapOrganizationToProvider2(org),
-        source: "swe-bench",
-        dimension: "coding",
-        score: Math.max(0, Math.min(100, entry.resolved_rate)),
-        rawScore: entry.resolved_rate,
-        metadata: {
-          organization: entry.organization,
-          date: entry.date
-        },
-        measuredAt: entry.date ? new Date(entry.date) : now,
-        ingestedAt: now
-      };
-    });
-    const { fromDate, toDate } = options ?? {};
-    if (fromDate) {
-      results = results.filter((r) => r.measuredAt >= fromDate);
-    }
-    if (toDate) {
-      results = results.filter((r) => r.measuredAt <= toDate);
-    }
-    return results;
-  }
-};
-function mapOrganizationToProvider2(org) {
-  const normalized = org.toLowerCase();
-  if (normalized.includes("openai"))
-    return "openai";
-  if (normalized.includes("anthropic"))
-    return "anthropic";
-  if (normalized.includes("google") || normalized.includes("deepmind"))
-    return "gemini";
-  if (normalized.includes("mistral"))
-    return "mistral";
-  if (normalized.includes("meta"))
-    return "meta";
-  return org;
-}
-
 // src/ingesters/open-llm-leaderboard.ts
 var DEFAULT_HF_URL = "https://huggingface.co/api/spaces/open-llm-leaderboard/open_llm_leaderboard/results";
 var BENCHMARK_MAPPINGS = [
@@ -344,7 +277,7 @@ var openLlmLeaderboardIngester = {
     for (const entry of entries) {
       const modelId = entry.model_name.toLowerCase().replace(/\s+/g, "-");
       const org = entry.organization?.toLowerCase() ?? "unknown";
-      const providerKey = mapOrganizationToProvider3(org);
+      const providerKey = mapOrganizationToProvider2(org);
       for (const mapping of BENCHMARK_MAPPINGS) {
         if (dims && !dims.has(mapping.dimension))
           continue;
@@ -371,6 +304,73 @@ var openLlmLeaderboardIngester = {
     return options?.maxResults ? results.slice(0, options.maxResults) : results;
   }
 };
+function mapOrganizationToProvider2(org) {
+  const normalized = org.toLowerCase();
+  if (normalized.includes("openai"))
+    return "openai";
+  if (normalized.includes("anthropic"))
+    return "anthropic";
+  if (normalized.includes("google") || normalized.includes("deepmind"))
+    return "gemini";
+  if (normalized.includes("mistral"))
+    return "mistral";
+  if (normalized.includes("meta"))
+    return "meta";
+  return org;
+}
+
+// src/ingesters/swe-bench.ts
+var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
+var sweBenchIngester = {
+  source: "swe-bench",
+  displayName: "SWE-bench",
+  description: "Software engineering task completion rates from SWE-bench.",
+  async ingest(options) {
+    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
+      return [];
+    }
+    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
+    const response = await fetchWithRetry(url, { fetch: options?.fetch });
+    const text = await response.text();
+    const data = parseJsonSafe(text, "SWE-bench");
+    const now = new Date;
+    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
+    if (options?.modelFilter?.length) {
+      const filterSet = new Set(options.modelFilter);
+      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
+    }
+    if (options?.maxResults) {
+      entries = entries.slice(0, options.maxResults);
+    }
+    let results = entries.map((entry) => {
+      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
+      const org = entry.organization?.toLowerCase() ?? "unknown";
+      return {
+        id: `swe-bench:${modelId}:coding`,
+        modelId,
+        providerKey: mapOrganizationToProvider3(org),
+        source: "swe-bench",
+        dimension: "coding",
+        score: Math.max(0, Math.min(100, entry.resolved_rate)),
+        rawScore: entry.resolved_rate,
+        metadata: {
+          organization: entry.organization,
+          date: entry.date
+        },
+        measuredAt: entry.date ? new Date(entry.date) : now,
+        ingestedAt: now
+      };
+    });
+    const { fromDate, toDate } = options ?? {};
+    if (fromDate) {
+      results = results.filter((r) => r.measuredAt >= fromDate);
+    }
+    if (toDate) {
+      results = results.filter((r) => r.measuredAt <= toDate);
+    }
+    return results;
+  }
+};
 function mapOrganizationToProvider3(org) {
   const normalized = org.toLowerCase();
   if (normalized.includes("openai"))

package/dist/node/ingesters/registry.js

@@ -249,73 +249,6 @@ function mapOrganizationToProvider(org) {
   return org;
 }
 
-// src/ingesters/swe-bench.ts
-var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
-var sweBenchIngester = {
-  source: "swe-bench",
-  displayName: "SWE-bench",
-  description: "Software engineering task completion rates from SWE-bench.",
-  async ingest(options) {
-    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
-      return [];
-    }
-    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
-    const response = await fetchWithRetry(url, { fetch: options?.fetch });
-    const text = await response.text();
-    const data = parseJsonSafe(text, "SWE-bench");
-    const now = new Date;
-    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
-    if (options?.modelFilter?.length) {
-      const filterSet = new Set(options.modelFilter);
-      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
-    }
-    if (options?.maxResults) {
-      entries = entries.slice(0, options.maxResults);
-    }
-    let results = entries.map((entry) => {
-      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
-      const org = entry.organization?.toLowerCase() ?? "unknown";
-      return {
-        id: `swe-bench:${modelId}:coding`,
-        modelId,
-        providerKey: mapOrganizationToProvider2(org),
-        source: "swe-bench",
-        dimension: "coding",
-        score: Math.max(0, Math.min(100, entry.resolved_rate)),
-        rawScore: entry.resolved_rate,
-        metadata: {
-          organization: entry.organization,
-          date: entry.date
-        },
-        measuredAt: entry.date ? new Date(entry.date) : now,
-        ingestedAt: now
-      };
-    });
-    const { fromDate, toDate } = options ?? {};
-    if (fromDate) {
-      results = results.filter((r) => r.measuredAt >= fromDate);
-    }
-    if (toDate) {
-      results = results.filter((r) => r.measuredAt <= toDate);
-    }
-    return results;
-  }
-};
-function mapOrganizationToProvider2(org) {
-  const normalized = org.toLowerCase();
-  if (normalized.includes("openai"))
-    return "openai";
-  if (normalized.includes("anthropic"))
-    return "anthropic";
-  if (normalized.includes("google") || normalized.includes("deepmind"))
-    return "gemini";
-  if (normalized.includes("mistral"))
-    return "mistral";
-  if (normalized.includes("meta"))
-    return "meta";
-  return org;
-}
-
 // src/ingesters/open-llm-leaderboard.ts
 var DEFAULT_HF_URL = "https://huggingface.co/api/spaces/open-llm-leaderboard/open_llm_leaderboard/results";
 var BENCHMARK_MAPPINGS = [
@@ -344,7 +277,7 @@ var openLlmLeaderboardIngester = {
     for (const entry of entries) {
       const modelId = entry.model_name.toLowerCase().replace(/\s+/g, "-");
       const org = entry.organization?.toLowerCase() ?? "unknown";
-      const providerKey = mapOrganizationToProvider3(org);
+      const providerKey = mapOrganizationToProvider2(org);
       for (const mapping of BENCHMARK_MAPPINGS) {
         if (dims && !dims.has(mapping.dimension))
           continue;
@@ -371,6 +304,73 @@ var openLlmLeaderboardIngester = {
     return options?.maxResults ? results.slice(0, options.maxResults) : results;
   }
 };
+function mapOrganizationToProvider2(org) {
+  const normalized = org.toLowerCase();
+  if (normalized.includes("openai"))
+    return "openai";
+  if (normalized.includes("anthropic"))
+    return "anthropic";
+  if (normalized.includes("google") || normalized.includes("deepmind"))
+    return "gemini";
+  if (normalized.includes("mistral"))
+    return "mistral";
+  if (normalized.includes("meta"))
+    return "meta";
+  return org;
+}
+
+// src/ingesters/swe-bench.ts
+var DEFAULT_SWE_BENCH_URL = "https://raw.githubusercontent.com/princeton-nlp/SWE-bench/main/docs/leaderboard.json";
+var sweBenchIngester = {
+  source: "swe-bench",
+  displayName: "SWE-bench",
+  description: "Software engineering task completion rates from SWE-bench.",
+  async ingest(options) {
+    if (options?.dimensions?.length && !options.dimensions.includes("coding")) {
+      return [];
+    }
+    const url = options?.sourceUrl ?? DEFAULT_SWE_BENCH_URL;
+    const response = await fetchWithRetry(url, { fetch: options?.fetch });
+    const text = await response.text();
+    const data = parseJsonSafe(text, "SWE-bench");
+    const now = new Date;
+    let entries = data.filter((entry) => entry.model && entry.resolved_rate != null);
+    if (options?.modelFilter?.length) {
+      const filterSet = new Set(options.modelFilter);
+      entries = entries.filter((e) => filterSet.has(e.model.toLowerCase().replace(/\s+/g, "-")));
+    }
+    if (options?.maxResults) {
+      entries = entries.slice(0, options.maxResults);
+    }
+    let results = entries.map((entry) => {
+      const modelId = entry.model.toLowerCase().replace(/\s+/g, "-");
+      const org = entry.organization?.toLowerCase() ?? "unknown";
+      return {
+        id: `swe-bench:${modelId}:coding`,
+        modelId,
+        providerKey: mapOrganizationToProvider3(org),
+        source: "swe-bench",
+        dimension: "coding",
+        score: Math.max(0, Math.min(100, entry.resolved_rate)),
+        rawScore: entry.resolved_rate,
+        metadata: {
+          organization: entry.organization,
+          date: entry.date
+        },
+        measuredAt: entry.date ? new Date(entry.date) : now,
+        ingestedAt: now
+      };
+    });
+    const { fromDate, toDate } = options ?? {};
+    if (fromDate) {
+      results = results.filter((r) => r.measuredAt >= fromDate);
+    }
+    if (toDate) {
+      results = results.filter((r) => r.measuredAt <= toDate);
+    }
+    return results;
+  }
+};
 function mapOrganizationToProvider3(org) {
   const normalized = org.toLowerCase();
   if (normalized.includes("openai"))

package/dist/scoring/index.d.ts

@@ -1,3 +1,3 @@
 export { computeModelRankings } from './composite-scorer';
 export { DEFAULT_DIMENSION_WEIGHTS, getWeightMap, normalizeWeights, } from './dimension-weights';
-export { normalizeScore, normalizeBenchmarkResults } from './normalizer';
+export { normalizeBenchmarkResults, normalizeScore } from './normalizer';

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@contractspec/lib.provider-ranking",
-  "version": "0.7.5",
+  "version": "0.7.8",
   "description": "AI provider ranking: benchmark ingestion, scoring, and model comparison",
   "keywords": [
     "contractspec",
@@ -24,16 +24,16 @@
     "dev": "contractspec-bun-build dev",
     "clean": "rimraf dist .turbo",
     "lint": "bun lint:fix",
-    "lint:fix": "eslint src --fix",
-    "lint:check": "eslint src",
+    "lint:fix": "biome check --write --unsafe --only=nursery/useSortedClasses . && biome check --write .",
+    "lint:check": "biome check .",
     "test": "bun test --pass-with-no-tests",
     "prebuild": "contractspec-bun-build prebuild",
     "typecheck": "tsc --noEmit"
   },
   "dependencies": {},
   "devDependencies": {
-    "@contractspec/tool.typescript": "3.7.5",
-    "@contractspec/tool.bun": "3.7.5",
+    "@contractspec/tool.typescript": "3.7.8",
+    "@contractspec/tool.bun": "3.7.8",
     "typescript": "^5.9.3"
   },
   "exports": {