@lunora/ai 1.0.0-alpha.65 → 1.0.0-alpha.66

package/dist/index.d.mts

@@ -63,9 +63,14 @@ interface ModelUsage {
  * An indicative price table, keyed by model id.
  *
  * Deliberately small: a table that tries to cover every model is a table that
- * is wrong about most of them. It holds the Workers AI models Lunora's own
- * defaults reference plus the common OpenAI embedding models, and everything
- * else returns `undefined` rather than a guess.
+ * is wrong about most of them. It holds the models Lunora's own defaults and
+ * documented examples reference — text generation and embeddings, on Workers AI
+ * and OpenAI — and everything else returns `undefined` rather than a guess.
+ *
+ * Generation models carry an `output` price; embedding models do not (they have
+ * no completion). A generation model missing from here is why a chat span would
+ * carry no cost at all off an AI Gateway, so the ids the docs teach are the ones
+ * that have to be in the table.
  */
 declare const DEFAULT_MODEL_PRICES: Readonly<Record<string, ModelPrice>>;
 /**

package/dist/index.d.ts

@@ -63,9 +63,14 @@ interface ModelUsage {
  * An indicative price table, keyed by model id.
  *
  * Deliberately small: a table that tries to cover every model is a table that
- * is wrong about most of them. It holds the Workers AI models Lunora's own
- * defaults reference plus the common OpenAI embedding models, and everything
- * else returns `undefined` rather than a guess.
+ * is wrong about most of them. It holds the models Lunora's own defaults and
+ * documented examples reference — text generation and embeddings, on Workers AI
+ * and OpenAI — and everything else returns `undefined` rather than a guess.
+ *
+ * Generation models carry an `output` price; embedding models do not (they have
+ * no completion). A generation model missing from here is why a chat span would
+ * carry no cost at all off an AI Gateway, so the ids the docs teach are the ones
+ * that have to be in the table.
  */
 declare const DEFAULT_MODEL_PRICES: Readonly<Record<string, ModelPrice>>;
 /**

package/dist/index.mjs

@@ -1 +1 @@
-import{default as o}from"./packem_shared/createAi-DIJS_f7H.mjs";import{AI_GATEWAY_ACCOUNT_ID_ENV as r,AI_GATEWAY_ID_ENV as A,AI_GATEWAY_TOKEN_ENV as _,buildAiGatewayMetadataFields as l,resolveAiGateway as m}from"./packem_shared/AI_GATEWAY_ACCOUNT_ID_ENV-9ptmPAYP.mjs";import{DEFAULT_MODEL_PRICES as E,estimateModelCost as T,lookupModelPrice as d}from"./packem_shared/DEFAULT_MODEL_PRICES-DZ3dYVUn.mjs";import{embed as x,embedMany as I,generateObject as c,generateText as f,hasToolCall as p,jsonSchema as b,streamObject as C,streamText as G,tool as M}from"ai";import{createWorkersAI as O}from"workers-ai-provider";export{r as AI_GATEWAY_ACCOUNT_ID_ENV,A as AI_GATEWAY_ID_ENV,_ as AI_GATEWAY_TOKEN_ENV,E as DEFAULT_MODEL_PRICES,l as buildAiGatewayMetadataFields,o as createAi,O as createWorkersAI,x as embed,I as embedMany,T as estimateModelCost,c as generateObject,f as generateText,p as hasToolCall,b as jsonSchema,d as lookupModelPrice,m as resolveAiGateway,C as streamObject,G as streamText,M as tool};
+import{default as o}from"./packem_shared/createAi-DIJS_f7H.mjs";import{AI_GATEWAY_ACCOUNT_ID_ENV as r,AI_GATEWAY_ID_ENV as A,AI_GATEWAY_TOKEN_ENV as _,buildAiGatewayMetadataFields as l,resolveAiGateway as m}from"./packem_shared/AI_GATEWAY_ACCOUNT_ID_ENV-9ptmPAYP.mjs";import{DEFAULT_MODEL_PRICES as E,estimateModelCost as T,lookupModelPrice as d}from"./packem_shared/DEFAULT_MODEL_PRICES-Q8uxdiuV.mjs";import{embed as x,embedMany as I,generateObject as c,generateText as f,hasToolCall as p,jsonSchema as b,streamObject as C,streamText as G,tool as M}from"ai";import{createWorkersAI as O}from"workers-ai-provider";export{r as AI_GATEWAY_ACCOUNT_ID_ENV,A as AI_GATEWAY_ID_ENV,_ as AI_GATEWAY_TOKEN_ENV,E as DEFAULT_MODEL_PRICES,l as buildAiGatewayMetadataFields,o as createAi,O as createWorkersAI,x as embed,I as embedMany,T as estimateModelCost,c as generateObject,f as generateText,p as hasToolCall,b as jsonSchema,d as lookupModelPrice,m as resolveAiGateway,C as streamObject,G as streamText,M as tool};

package/dist/packem_shared/DEFAULT_MODEL_PRICES-Q8uxdiuV.mjs

@@ -0,0 +1 @@
+const p=/^(.*)-\d{4}-\d{2}-\d{2}$/u,a={"@cf/baai/bge-base-en-v1.5":{input:.067},"@cf/baai/bge-large-en-v1.5":{input:.204},"@cf/baai/bge-m3":{input:.012},"@cf/baai/bge-small-en-v1.5":{input:.02},"@cf/meta/llama-3.1-8b-instruct":{input:.28,output:.83},"@cf/meta/llama-3.3-70b-instruct-fp8-fast":{input:.29,output:2.25},"text-embedding-3-large":{input:.13},"text-embedding-3-small":{input:.02},"gpt-4o":{input:2.5,output:10},"gpt-4o-mini":{input:.15,output:.6},"gpt-5":{input:1.25,output:10}},r=n=>{const t=n.trim(),i=t.lastIndexOf("/");return(i!==-1&&!t.startsWith("@")?[t,t.slice(i+1)]:[t]).flatMap(e=>{const o=p.exec(e)?.[1];return o===void 0?[e]:[e,o]})},c=(n,t=a)=>{for(const i of r(n))if(Object.hasOwn(t,i))return t[i]},b=(n,t,i)=>{if(n===void 0||n.length===0)return;const u=c(n,i);if(u===void 0)return;const e=Number.isFinite(t.inputTokens)?Math.max(0,t.inputTokens):0,o=Number.isFinite(t.outputTokens)?Math.max(0,t.outputTokens):0;if(e<=0&&o<=0)return;const s=(e*u.input+o*(u.output??0))/1e6;return Number.isFinite(s)?s:void 0};export{a as DEFAULT_MODEL_PRICES,b as estimateModelCost,c as lookupModelPrice};

package/dist/packem_shared/{defineRag-Dzh-MbCu.mjs → defineRag-CMM17YeT.mjs}

@@ -1,4 +1,4 @@
-import{LunoraError as g,isLunoraError as ye}from"@lunora/errors";import{tool as xe,jsonSchema as Ie,embedMany as Se,embed as _e}from"ai";import{estimateModelCost as ke}from"./DEFAULT_MODEL_PRICES-DZ3dYVUn.mjs";import Te from"./fixedWindowChunks-C461ahRE.mjs";import{c as Ae,I as Ne}from"./concurrent-C6nqBv41.mjs";import{contentHash as Me}from"./contentHash-BIn6ECP8.mjs";import ee from"./hybridRank-DejmVw2I.mjs";import{VECTORIZE_CAPABILITIES as te,vectorizeStore as De}from"./VECTORIZE_CAPABILITIES-CUQDoxis.mjs";const Re=1e3,Ce=200,$e=5,Be=4,ne=te.maxMetadataBytes===!1?Number.POSITIVE_INFINITY:te.maxMetadataBytes,Oe=2*1024,ie="__ragChunk",ce="__ragSource",A="__ragText",Q="__ragHash",V="__ragChunks",B="__ragImportance",de="__ragModel",Ke=new Set([ie,V,Q,B,de,ce,A]),Ue=(e,i,m,w)=>{if(w===!1)return;const E=new TextEncoder().encode(JSON.stringify(e)).length;if(E<=w)return;const R=(typeof e[A]=="string"?new TextEncoder().encode(e[A]).length:0)*2>E?"lower `chunkSize` (it counts characters, not bytes — multibyte text costs up to 3 bytes each), or supply `textStore` to move chunk text out of metadata entirely":"attach less per-source `metadata`";throw new g("BAD_REQUEST",`@lunora/ai/rag: chunk ${String(i)} of "${m}" carries ${String(E)} bytes of metadata, over the store's ${String(w)}-byte per-vector ceiling — ${R}`)},Qe=(e,i,m)=>{if(m===!1)return;const w=new TextEncoder().encode(e).length;if(!(w<=m))throw new g("BAD_REQUEST",`@lunora/ai/rag: chunk id "${e}" for source "${i}" is ${String(w)} bytes, over the store's ${String(m)}-byte per-vector id ceiling — shorten the source id (hash long keys before indexing them) or shorten the \`namespace\`, which is prefixed onto every chunk id`)},Ve=/^[\w.-]{1,40}$/,ue=e=>e===void 0?"":`${encodeURIComponent(e)}#`,T=(e,i,m)=>`${ue(e)}${i}#${String(m)}`,re=(e,i)=>{const m=ue(i),w=m!==""&&e.startsWith(m)?e.slice(m.length):e,E=w.lastIndexOf("#"),S=E===-1?Number.NaN:Number(w.slice(E+1));return E===-1||!Number.isInteger(S)||S<0?{chunkIndex:0,sourceId:w}:{chunkIndex:S,sourceId:w.slice(0,E)}},Fe=async e=>Me(new TextEncoder().encode(e)),ze=(e,i)=>e.filter(m=>m.score>=i),oe=e=>{if(!e)return;const i=Object.entries(e).filter(([m])=>!Ke.has(m));return i.length>0?Object.fromEntries(i):void 0},se=new Set,Le=e=>{se.has(e)||(se.add(e),console.warn(`[@lunora/ai/rag] index "${e}" is used without a namespace — in a multi-tenant/sharded
+import{LunoraError as g,isLunoraError as ye}from"@lunora/errors";import{tool as xe,jsonSchema as Ie,embedMany as Se,embed as _e}from"ai";import{estimateModelCost as ke}from"./DEFAULT_MODEL_PRICES-Q8uxdiuV.mjs";import Te from"./fixedWindowChunks-C461ahRE.mjs";import{c as Ae,I as Ne}from"./concurrent-C6nqBv41.mjs";import{contentHash as Me}from"./contentHash-BIn6ECP8.mjs";import ee from"./hybridRank-DejmVw2I.mjs";import{VECTORIZE_CAPABILITIES as te,vectorizeStore as De}from"./VECTORIZE_CAPABILITIES-CUQDoxis.mjs";const Re=1e3,Ce=200,$e=5,Be=4,ne=te.maxMetadataBytes===!1?Number.POSITIVE_INFINITY:te.maxMetadataBytes,Oe=2*1024,ie="__ragChunk",ce="__ragSource",A="__ragText",Q="__ragHash",V="__ragChunks",B="__ragImportance",de="__ragModel",Ke=new Set([ie,V,Q,B,de,ce,A]),Ue=(e,i,m,w)=>{if(w===!1)return;const E=new TextEncoder().encode(JSON.stringify(e)).length;if(E<=w)return;const R=(typeof e[A]=="string"?new TextEncoder().encode(e[A]).length:0)*2>E?"lower `chunkSize` (it counts characters, not bytes — multibyte text costs up to 3 bytes each), or supply `textStore` to move chunk text out of metadata entirely":"attach less per-source `metadata`";throw new g("BAD_REQUEST",`@lunora/ai/rag: chunk ${String(i)} of "${m}" carries ${String(E)} bytes of metadata, over the store's ${String(w)}-byte per-vector ceiling — ${R}`)},Qe=(e,i,m)=>{if(m===!1)return;const w=new TextEncoder().encode(e).length;if(!(w<=m))throw new g("BAD_REQUEST",`@lunora/ai/rag: chunk id "${e}" for source "${i}" is ${String(w)} bytes, over the store's ${String(m)}-byte per-vector id ceiling — shorten the source id (hash long keys before indexing them) or shorten the \`namespace\`, which is prefixed onto every chunk id`)},Ve=/^[\w.-]{1,40}$/,ue=e=>e===void 0?"":`${encodeURIComponent(e)}#`,T=(e,i,m)=>`${ue(e)}${i}#${String(m)}`,re=(e,i)=>{const m=ue(i),w=m!==""&&e.startsWith(m)?e.slice(m.length):e,E=w.lastIndexOf("#"),S=E===-1?Number.NaN:Number(w.slice(E+1));return E===-1||!Number.isInteger(S)||S<0?{chunkIndex:0,sourceId:w}:{chunkIndex:S,sourceId:w.slice(0,E)}},Fe=async e=>Me(new TextEncoder().encode(e)),ze=(e,i)=>e.filter(m=>m.score>=i),oe=e=>{if(!e)return;const i=Object.entries(e).filter(([m])=>!Ke.has(m));return i.length>0?Object.fromEntries(i):void 0},se=new Set,Le=e=>{se.has(e)||(se.add(e),console.warn(`[@lunora/ai/rag] index "${e}" is used without a namespace — in a multi-tenant/sharded
 app this shares one tenant's chunks (text included) with every other tenant, since
 Vectorize indexes are account-global. Pass \`namespace\` (the shard/tenant key) on both
 index() and retrieve(). Single-tenant apps suppress this via { allowSharedNamespace: true }.`))},Pe=e=>e.map(i=>`[source:${i.sourceId}#${String(i.chunkIndex)}]

package/dist/rag/index.mjs

@@ -1 +1 @@
-import{default as o}from"../packem_shared/fixedWindowChunks-C461ahRE.mjs";import{markdownChunker as a,sentenceChunker as f,tokenChunker as n}from"../packem_shared/markdownChunker-CDME5gEs.mjs";import{default as x}from"../packem_shared/defineRag-Dzh-MbCu.mjs";import{contentHash as p,guessMimeTypeFromExtension as i}from"../packem_shared/contentHash-BIn6ECP8.mjs";import{default as c}from"../packem_shared/hybridRank-DejmVw2I.mjs";import{default as l}from"../packem_shared/bm25LexicalStore-CALA7Lhm.mjs";import{default as h}from"../packem_shared/matchesMetadataFilter-BbIOyA5g.mjs";import{batchReranker as g,scoreReranker as C}from"../packem_shared/batchReranker-Bc38FBLH.mjs";import{defineRagSource as E}from"../packem_shared/defineRagSource-Q3f3niU8.mjs";import{sqlLexicalStore as T}from"../packem_shared/sqlLexicalStore-qkTJHHzM.mjs";import{sqliteVectorStore as y}from"../packem_shared/sqliteVectorStore-DEqrp-Uc.mjs";import{ragSyncTriggers as q}from"../packem_shared/ragSyncTriggers-DPqzBNFw.mjs";import{VECTORIZE_CAPABILITIES as A,vectorizeStore as F}from"../packem_shared/VECTORIZE_CAPABILITIES-CUQDoxis.mjs";export{A as VECTORIZE_CAPABILITIES,g as batchReranker,l as bm25LexicalStore,p as contentHash,x as defineRag,E as defineRagSource,o as fixedWindowChunks,i as guessMimeTypeFromExtension,c as hybridRank,a as markdownChunker,h as matchesMetadataFilter,q as ragSyncTriggers,C as scoreReranker,f as sentenceChunker,T as sqlLexicalStore,y as sqliteVectorStore,n as tokenChunker,F as vectorizeStore};
+import{default as o}from"../packem_shared/fixedWindowChunks-C461ahRE.mjs";import{markdownChunker as a,sentenceChunker as f,tokenChunker as n}from"../packem_shared/markdownChunker-CDME5gEs.mjs";import{default as x}from"../packem_shared/defineRag-CMM17YeT.mjs";import{contentHash as p,guessMimeTypeFromExtension as i}from"../packem_shared/contentHash-BIn6ECP8.mjs";import{default as c}from"../packem_shared/hybridRank-DejmVw2I.mjs";import{default as l}from"../packem_shared/bm25LexicalStore-CALA7Lhm.mjs";import{default as h}from"../packem_shared/matchesMetadataFilter-BbIOyA5g.mjs";import{batchReranker as g,scoreReranker as C}from"../packem_shared/batchReranker-Bc38FBLH.mjs";import{defineRagSource as E}from"../packem_shared/defineRagSource-Q3f3niU8.mjs";import{sqlLexicalStore as T}from"../packem_shared/sqlLexicalStore-qkTJHHzM.mjs";import{sqliteVectorStore as y}from"../packem_shared/sqliteVectorStore-DEqrp-Uc.mjs";import{ragSyncTriggers as q}from"../packem_shared/ragSyncTriggers-DPqzBNFw.mjs";import{VECTORIZE_CAPABILITIES as A,vectorizeStore as F}from"../packem_shared/VECTORIZE_CAPABILITIES-CUQDoxis.mjs";export{A as VECTORIZE_CAPABILITIES,g as batchReranker,l as bm25LexicalStore,p as contentHash,x as defineRag,E as defineRagSource,o as fixedWindowChunks,i as guessMimeTypeFromExtension,c as hybridRank,a as markdownChunker,h as matchesMetadataFilter,q as ragSyncTriggers,C as scoreReranker,f as sentenceChunker,T as sqlLexicalStore,y as sqliteVectorStore,n as tokenChunker,F as vectorizeStore};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lunora/ai",
-  "version": "1.0.0-alpha.65",
+  "version": "1.0.0-alpha.66",
   "description": "Workers AI helper for Lunora: provider-agnostic AI SDK access from functions, Workers AI by default",
   "keywords": [
     "ai",

package/dist/packem_shared/DEFAULT_MODEL_PRICES-DZ3dYVUn.mjs

@@ -1 +0,0 @@
-const r=/^(.*)-\d{4}-\d{2}-\d{2}$/u,a={"@cf/baai/bge-base-en-v1.5":{input:.067},"@cf/baai/bge-large-en-v1.5":{input:.204},"@cf/baai/bge-m3":{input:.012},"@cf/baai/bge-small-en-v1.5":{input:.02},"text-embedding-3-large":{input:.13},"text-embedding-3-small":{input:.02}},c=n=>{const t=n.trim(),e=t.lastIndexOf("/");return(e!==-1&&!t.startsWith("@")?[t,t.slice(e+1)]:[t]).flatMap(i=>{const o=r.exec(i)?.[1];return o===void 0?[i]:[i,o]})},b=(n,t=a)=>{for(const e of c(n))if(Object.hasOwn(t,e))return t[e]},d=(n,t,e)=>{if(n===void 0||n.length===0)return;const s=b(n,e);if(s===void 0)return;const i=Number.isFinite(t.inputTokens)?Math.max(0,t.inputTokens):0,o=Number.isFinite(t.outputTokens)?Math.max(0,t.outputTokens):0;if(i<=0&&o<=0)return;const u=(i*s.input+o*(s.output??0))/1e6;return Number.isFinite(u)?u:void 0};export{a as DEFAULT_MODEL_PRICES,d as estimateModelCost,b as lookupModelPrice};