SPB Git

spb/llmindex Public

The discriminative, contamination-resistant, fully transparent LLM ranking — updated live.

TypeScript 77.9% TeX 15.2% Python 3.7% SQL 1.4% JavaScript 1.1% Shell 0.5%
5.3 KB · 156 lines sql
Raw Blame History
1-- CreateTable2CREATE TABLE "models" (3    "id" TEXT NOT NULL,4    "slug" TEXT NOT NULL,5    "name" TEXT NOT NULL,6    "provider" TEXT NOT NULL,7    "context_length" INTEGER,8    "prompt_price_per_m" DOUBLE PRECISION,9    "completion_price_per_m" DOUBLE PRECISION,10    "active" BOOLEAN NOT NULL DEFAULT true,11    "ranked" BOOLEAN NOT NULL DEFAULT false,12    "synced_at" TIMESTAMP(3) NOT NULL,13    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,1415    CONSTRAINT "models_pkey" PRIMARY KEY ("id")16);1718-- CreateTable19CREATE TABLE "eval_items" (20    "id" TEXT NOT NULL,21    "template_id" TEXT NOT NULL,22    "domain" TEXT NOT NULL,23    "prompt" TEXT NOT NULL,24    "answer_key" TEXT NOT NULL,25    "rubric" TEXT,26    "perturb_seed" TEXT,27    "is_anchor" BOOLEAN NOT NULL DEFAULT false,28    "irt_a" DOUBLE PRECISION,29    "irt_b" DOUBLE PRECISION,30    "status" TEXT NOT NULL DEFAULT 'active',31    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,3233    CONSTRAINT "eval_items_pkey" PRIMARY KEY ("id")34);3536-- CreateTable37CREATE TABLE "score_runs" (38    "id" TEXT NOT NULL,39    "index_version" TEXT NOT NULL,40    "kind" TEXT NOT NULL,41    "status" TEXT NOT NULL DEFAULT 'pending',42    "item_set_hash" TEXT NOT NULL,43    "model_set" JSONB NOT NULL,44    "fit_diagnostics" JSONB,45    "notes" TEXT,46    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,47    "completed_at" TIMESTAMP(3),4849    CONSTRAINT "score_runs_pkey" PRIMARY KEY ("id")50);5152-- CreateTable53CREATE TABLE "model_responses" (54    "id" TEXT NOT NULL,55    "run_id" TEXT NOT NULL,56    "model_id" TEXT NOT NULL,57    "item_id" TEXT NOT NULL,58    "sample_index" INTEGER NOT NULL DEFAULT 0,59    "request_params" JSONB NOT NULL,60    "raw_response" JSONB NOT NULL,61    "answer_extracted" TEXT,62    "correct" BOOLEAN,63    "confidence" DOUBLE PRECISION,64    "tokens_in" INTEGER,65    "tokens_out" INTEGER,66    "latency_ms" INTEGER,67    "cost_usd" DOUBLE PRECISION,68    "error" TEXT,69    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,7071    CONSTRAINT "model_responses_pkey" PRIMARY KEY ("id")72);7374-- CreateTable75CREATE TABLE "pairwise_duels" (76    "id" TEXT NOT NULL,77    "run_id" TEXT NOT NULL,78    "domain" TEXT NOT NULL,79    "item_id" TEXT NOT NULL,80    "model_a_id" TEXT NOT NULL,81    "model_b_id" TEXT NOT NULL,82    "judge_slug" TEXT NOT NULL,83    "position_swapped" BOOLEAN NOT NULL DEFAULT false,84    "winner" TEXT NOT NULL,85    "raw_judgment" JSONB NOT NULL,86    "created_at" TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP,8788    CONSTRAINT "pairwise_duels_pkey" PRIMARY KEY ("id")89);9091-- CreateTable92CREATE TABLE "scores" (93    "id" TEXT NOT NULL,94    "run_id" TEXT NOT NULL,95    "model_id" TEXT NOT NULL,96    "domain" TEXT NOT NULL,97    "score" DOUBLE PRECISION NOT NULL,98    "score_low" DOUBLE PRECISION NOT NULL,99    "score_high" DOUBLE PRECISION NOT NULL,100    "sub_metrics" JSONB,101102    CONSTRAINT "scores_pkey" PRIMARY KEY ("id")103);104105-- CreateIndex106CREATE UNIQUE INDEX "models_slug_key" ON "models"("slug");107108-- CreateIndex109CREATE INDEX "eval_items_domain_status_idx" ON "eval_items"("domain", "status");110111-- CreateIndex112CREATE INDEX "score_runs_kind_status_created_at_idx" ON "score_runs"("kind", "status", "created_at");113114-- CreateIndex115CREATE INDEX "model_responses_model_id_created_at_idx" ON "model_responses"("model_id", "created_at");116117-- CreateIndex118CREATE UNIQUE INDEX "model_responses_run_id_model_id_item_id_sample_index_key" ON "model_responses"("run_id", "model_id", "item_id", "sample_index");119120-- CreateIndex121CREATE INDEX "pairwise_duels_run_id_domain_idx" ON "pairwise_duels"("run_id", "domain");122123-- CreateIndex124CREATE INDEX "scores_domain_idx" ON "scores"("domain");125126-- CreateIndex127CREATE UNIQUE INDEX "scores_run_id_model_id_domain_key" ON "scores"("run_id", "model_id", "domain");128129-- AddForeignKey130ALTER TABLE "model_responses" ADD CONSTRAINT "model_responses_run_id_fkey" FOREIGN KEY ("run_id") REFERENCES "score_runs"("id") ON DELETE RESTRICT ON UPDATE CASCADE;131132-- AddForeignKey133ALTER TABLE "model_responses" ADD CONSTRAINT "model_responses_model_id_fkey" FOREIGN KEY ("model_id") REFERENCES "models"("id") ON DELETE RESTRICT ON UPDATE CASCADE;134135-- AddForeignKey136ALTER TABLE "model_responses" ADD CONSTRAINT "model_responses_item_id_fkey" FOREIGN KEY ("item_id") REFERENCES "eval_items"("id") ON DELETE RESTRICT ON UPDATE CASCADE;137138-- AddForeignKey139ALTER TABLE "pairwise_duels" ADD CONSTRAINT "pairwise_duels_run_id_fkey" FOREIGN KEY ("run_id") REFERENCES "score_runs"("id") ON DELETE RESTRICT ON UPDATE CASCADE;140141-- AddForeignKey142ALTER TABLE "pairwise_duels" ADD CONSTRAINT "pairwise_duels_item_id_fkey" FOREIGN KEY ("item_id") REFERENCES "eval_items"("id") ON DELETE RESTRICT ON UPDATE CASCADE;143144-- AddForeignKey145ALTER TABLE "pairwise_duels" ADD CONSTRAINT "pairwise_duels_model_a_id_fkey" FOREIGN KEY ("model_a_id") REFERENCES "models"("id") ON DELETE RESTRICT ON UPDATE CASCADE;146147-- AddForeignKey148ALTER TABLE "pairwise_duels" ADD CONSTRAINT "pairwise_duels_model_b_id_fkey" FOREIGN KEY ("model_b_id") REFERENCES "models"("id") ON DELETE RESTRICT ON UPDATE CASCADE;149150-- AddForeignKey151ALTER TABLE "scores" ADD CONSTRAINT "scores_run_id_fkey" FOREIGN KEY ("run_id") REFERENCES "score_runs"("id") ON DELETE RESTRICT ON UPDATE CASCADE;152153-- AddForeignKey154ALTER TABLE "scores" ADD CONSTRAINT "scores_model_id_fkey" FOREIGN KEY ("model_id") REFERENCES "models"("id") ON DELETE RESTRICT ON UPDATE CASCADE;155156