parsefy 1.0.2 → 1.0.4

package/README.md

@@ -45,8 +45,8 @@ if (!error && object) {
   console.log(object.invoice_number); // Fully typed!
 
   // Access field-level confidence and evidence
-  console.log(`Overall confidence: ${metadata.confidenceScore}`);
-  metadata.fieldConfidence.forEach((fc) => {
+  console.log(`Overall confidence: ${metadata.confidence_score}`);
+  metadata.field_confidence.forEach((fc) => {
     console.log(`${fc.field}: ${fc.score} (${fc.reason}) - "${fc.text}"`);
   });
 }
@@ -111,15 +111,15 @@ interface ExtractResult<T> {
 
   // Metadata about the extraction
   metadata: {
-    processingTimeMs: number;     // Processing time in milliseconds
-    inputTokens: number;          // Input tokens used
-    outputTokens: number;         // Output tokens generated
+    processing_time_ms: number;     // Processing time in milliseconds
+    input_tokens: number;          // Input tokens used
+    output_tokens: number;         // Output tokens generated
     credits: number;              // Credits consumed (1 credit = 1 page)
-    fallbackTriggered: boolean;   // Whether fallback model was used
+    fallback_triggered: boolean;   // Whether fallback model was used
 
     // 🆕 Field-level confidence and evidence
-    confidenceScore: number;      // Overall confidence (0.0 to 1.0)
-    fieldConfidence: Array<{
+    confidence_score: number;      // Overall confidence (0.0 to 1.0)
+    field_confidence: Array<{
       field: string;              // JSON path (e.g., "$.invoice_number")
       score: number;              // Confidence score (0.0 to 1.0)
       reason: string;             // "Exact match", "Inferred from header", etc.
@@ -150,9 +150,9 @@ const { object, metadata } = await client.extract({ file, schema });
   vendor: "Acme Corp"
 }
 
-// metadata.confidenceScore: 0.94
+// metadata.confidence_score: 0.94
 
-// metadata.fieldConfidence:
+// metadata.field_confidence:
 [
   { field: "$.invoice_number", score: 0.98, reason: "Exact match", page: 1, text: "Invoice # INV-2024-0042" },
   { field: "$.date", score: 0.95, reason: "Exact match", page: 1, text: "Date: 01/15/2024" },
@@ -286,8 +286,8 @@ app.post('/extract', upload.single('document'), async (req, res) => {
 
   res.json({
     data: object,
-    confidence: metadata.confidenceScore,
-    fieldDetails: metadata.fieldConfidence,
+    confidence: metadata.confidence_score,
+    fieldDetails: metadata.field_confidence,
     error,
   });
 });
@@ -313,7 +313,7 @@ app.post('/extract', async (c) => {
 
   return c.json({
     data: object,
-    confidence: metadata.confidenceScore,
+    confidence: metadata.confidence_score,
     issues: metadata.issues,
     error,
   });
@@ -334,13 +334,13 @@ try {
   // Extraction-level errors (request succeeded, but extraction failed)
   if (error) {
     console.error(`Extraction failed: [${error.code}] ${error.message}`);
-    console.log(`Fallback triggered: ${metadata.fallbackTriggered}`);
+    console.log(`Fallback triggered: ${metadata.fallback_triggered}`);
     console.log(`Issues: ${metadata.issues.join(', ')}`);
     return;
   }
 
   // Check for low confidence fields
-  const lowConfidence = metadata.fieldConfidence.filter((fc) => fc.score < 0.80);
+  const lowConfidence = metadata.field_confidence.filter((fc) => fc.score < 0.80);
   if (lowConfidence.length > 0) {
     console.warn('Low confidence fields:', lowConfidence);
   }

package/dist/index.cjs

@@ -1,2 +1,2 @@
-'use strict';var zodToJsonSchema=require('zod-to-json-schema');var u=.85,d={".pdf":"application/pdf",".docx":"application/vnd.openxmlformats-officedocument.wordprocessingml.document"},l=10*1024*1024,h="https://api.parsefy.io",E=6e4;var i=class extends Error{constructor(e,r){super(e),this.name="ParsefyError",this.code=r,typeof Error.captureStackTrace=="function"&&Error.captureStackTrace(this,this.constructor);}},p=class extends i{constructor(e,r,o){super(e),this.name="APIError",this.statusCode=r,this.response=o;}},y=class extends i{constructor(e,r,o){super(e,r),this.name="ExtractionError",this.metadata=o;}},s=class extends i{constructor(e){super(e),this.name="ValidationError";}};function x(){return typeof process<"u"&&process.versions?.node!==void 0}function b(t){let e=zodToJsonSchema.zodToJsonSchema(t,{$refStrategy:"none",target:"jsonSchema7"});return "$schema"in e&&delete e.$schema,e}function R(t){let e=t.toLowerCase().match(/\.[^.]+$/)?.[0];return e&&d[e]||null}function w(t){if(!R(t)){let r=Object.keys(d).join(", ");throw new s(`Unsupported file type. Supported types: ${r}`)}}function f(t){if(t===0)throw new s("File is empty");if(t>l){let e=l/1048576;throw new s(`File size exceeds maximum limit of ${e}MB`)}}function F(t){return {object:t.object,metadata:{processingTimeMs:t.metadata.processing_time_ms,inputTokens:t.metadata.input_tokens,outputTokens:t.metadata.output_tokens,credits:t.metadata.credits,fallbackTriggered:t.metadata.fallback_triggered,confidenceScore:t._meta.confidence_score,fieldConfidence:t._meta.field_confidence.map(e=>({field:e.field,score:e.score,reason:e.reason,page:e.page,text:e.text})),issues:t._meta.issues},error:t.error}}function T(t,e){let r=R(e)||"application/octet-stream",o=t.buffer.slice(t.byteOffset,t.byteOffset+t.byteLength);return typeof File<"u"?new File([o],e,{type:r}):new Blob([o],{type:r})}async function k(t){if(!x())throw new s("File path strings are only supported in Node.js. Use File or Blob in the browser.");let e=await import('fs'),r=await import('path');if(!e.existsSync(t))throw new s(`File not found: ${t}`);let o=r.basename(t);w(o);let a=e.readFileSync(t);return f(a.length),{buffer:a,filename:o}}async function _(t){if(typeof t=="string"){let{buffer:e,filename:r}=await k(t);return T(e,r)}if(Buffer.isBuffer(t))return f(t.length),T(t,"document.pdf");if(t instanceof File)return w(t.name),f(t.size),t;if(t instanceof Blob)return f(t.size),t;throw new s("Invalid file input. Expected File, Blob, Buffer, or file path string.")}function P(t){return new Promise(e=>setTimeout(e,t))}function S(t,e=1e3){let r=e*Math.pow(2,t),o=Math.random()*.1*r;return Math.min(r+o,3e4)}var g=class{constructor(e){this.maxRetries=3;let r={};if(typeof e=="string"?r={apiKey:e}:e&&(r=e),this.apiKey=r.apiKey||this.getEnvApiKey(),!this.apiKey)throw new s("API key is required. Provide it in the constructor or set the PARSEFY_API_KEY environment variable.");this.baseUrl=r.baseUrl||h,this.timeout=r.timeout||E;}getEnvApiKey(){return x()&&process.env.PARSEFY_API_KEY||""}async extract(e){let{file:r,schema:o,confidenceThreshold:a}=e,n=b(o),m=await _(r),c=new FormData;return c.append("file",m),c.append("output_schema",JSON.stringify(n)),c.append("confidence_threshold",String(a??.85)),this.makeRequestWithRetry(c)}async makeRequestWithRetry(e,r=0){try{return await this.makeRequest(e)}catch(o){if(o instanceof p&&o.statusCode===429&&r<this.maxRetries){let a=S(r);return await P(a),this.makeRequestWithRetry(e,r+1)}throw o}}async makeRequest(e){let r=`${this.baseUrl}/v1/extract`,o=new AbortController,a=setTimeout(()=>o.abort(),this.timeout);try{let n=await fetch(r,{method:"POST",headers:{Authorization:`Bearer ${this.apiKey}`},body:e,signal:o.signal});if(clearTimeout(a),!n.ok){let c=await this.parseErrorResponse(n);throw new p(c.message||`API request failed with status ${n.status}`,n.status,c)}let m=await n.json();return F(m)}catch(n){throw clearTimeout(a),n instanceof Error&&n.name==="AbortError"?new i(`Request timed out after ${this.timeout}ms`,"TIMEOUT"):n instanceof i?n:n instanceof TypeError?new i("Network error: Unable to connect to the Parsefy API","NETWORK_ERROR"):new i(`Unexpected error: ${n instanceof Error?n.message:String(n)}`,"UNKNOWN_ERROR")}}async parseErrorResponse(e){try{return await e.json()}catch{try{return {message:await e.text()||e.statusText}}catch{return {message:e.statusText}}}}};
-exports.APIError=p;exports.DEFAULT_CONFIDENCE_THRESHOLD=u;exports.ExtractionError=y;exports.Parsefy=g;exports.ParsefyError=i;exports.ValidationError=s;
+'use strict';var zodToJsonSchema=require('zod-to-json-schema');var u=.85,d={".pdf":"application/pdf",".docx":"application/vnd.openxmlformats-officedocument.wordprocessingml.document"},l=10*1024*1024,g="https://api.parsefy.io",x=6e4;var s=class extends Error{constructor(e,r){super(e),this.name="ParsefyError",this.code=r,typeof Error.captureStackTrace=="function"&&Error.captureStackTrace(this,this.constructor);}},p=class extends s{constructor(e,r,o){super(e),this.name="APIError",this.statusCode=r,this.response=o;}},y=class extends s{constructor(e,r,o){super(e,r),this.name="ExtractionError",this.metadata=o;}},a=class extends s{constructor(e){super(e),this.name="ValidationError";}};function h(){return typeof process<"u"&&process.versions?.node!==void 0}function T(t){let e=zodToJsonSchema.zodToJsonSchema(t,{$refStrategy:"none",target:"jsonSchema7"});return "$schema"in e&&delete e.$schema,e}function b(t){let e=t.toLowerCase().match(/\.[^.]+$/)?.[0];return e&&d[e]||null}function _(t){if(!b(t)){let r=Object.keys(d).join(", ");throw new a(`Unsupported file type. Supported types: ${r}`)}}function m(t){if(t===0)throw new a("File is empty");if(t>l){let e=l/1048576;throw new a(`File size exceeds maximum limit of ${e}MB`)}}function w(t){let e=t._meta||{confidence_score:1,field_confidence:[],issues:[]};return {object:t.object,metadata:{processing_time_ms:t.metadata.processing_time_ms,input_tokens:t.metadata.input_tokens,output_tokens:t.metadata.output_tokens,credits:t.metadata.credits,fallback_triggered:t.metadata.fallback_triggered,confidence_score:e.confidence_score,field_confidence:e.field_confidence.map(r=>({field:r.field,score:r.score,reason:r.reason,page:r.page,text:r.text})),issues:e.issues},error:t.error}}function R(t,e){let r=b(e)||"application/octet-stream",o=t.buffer.slice(t.byteOffset,t.byteOffset+t.byteLength);return typeof File<"u"?new File([o],e,{type:r}):new Blob([o],{type:r})}async function I(t){if(!h())throw new a("File path strings are only supported in Node.js. Use File or Blob in the browser.");let e=await import('fs'),r=await import('path');if(!e.existsSync(t))throw new a(`File not found: ${t}`);let o=r.basename(t);_(o);let c=e.readFileSync(t);return m(c.length),{buffer:c,filename:o}}async function F(t){if(typeof t=="string"){let{buffer:e,filename:r}=await I(t);return R(e,r)}if(Buffer.isBuffer(t))return m(t.length),R(t,"document.pdf");if(t instanceof File)return _(t.name),m(t.size),t;if(t instanceof Blob)return m(t.size),t;throw new a("Invalid file input. Expected File, Blob, Buffer, or file path string.")}function P(t){return new Promise(e=>setTimeout(e,t))}function A(t,e=1e3){let r=e*Math.pow(2,t),o=Math.random()*.1*r;return Math.min(r+o,3e4)}var E=class{constructor(e){this.maxRetries=3;let r={};if(typeof e=="string"?r={apiKey:e}:e&&(r=e),this.apiKey=r.apiKey||this.getEnvApiKey(),!this.apiKey)throw new a("API key is required. Provide it in the constructor or set the PARSEFY_API_KEY environment variable.");this.baseUrl=r.baseUrl||g,this.timeout=r.timeout||x;}getEnvApiKey(){return h()&&process.env.PARSEFY_API_KEY||""}async extract(e){let{file:r,schema:o,confidenceThreshold:c}=e,n=T(o),f=await F(r),i=new FormData;return i.append("file",f),i.append("output_schema",JSON.stringify(n)),i.append("confidence_threshold",String(c??.85)),this.makeRequestWithRetry(i)}async makeRequestWithRetry(e,r=0){try{return await this.makeRequest(e)}catch(o){if(o instanceof p&&o.statusCode===429&&r<this.maxRetries){let c=A(r);return await P(c),this.makeRequestWithRetry(e,r+1)}throw o}}async makeRequest(e){let r=`${this.baseUrl}/v1/extract`,o=new AbortController,c=setTimeout(()=>o.abort(),this.timeout);try{let n=await fetch(r,{method:"POST",headers:{Authorization:`Bearer ${this.apiKey}`},body:e,signal:o.signal});if(clearTimeout(c),!n.ok){let i=await this.parseErrorResponse(n);throw new p(i.message||`API request failed with status ${n.status}`,n.status,i)}let f;try{f=await n.json();}catch{throw new s("Failed to parse API response as JSON. The API may have returned an invalid response.","PARSE_ERROR")}try{return w(f)}catch(i){throw new s(`Failed to transform API response: ${i instanceof Error?i.message:String(i)}`,"TRANSFORM_ERROR")}}catch(n){throw clearTimeout(c),n instanceof Error&&n.name==="AbortError"?new s(`Request timed out after ${this.timeout}ms`,"TIMEOUT"):n instanceof s?n:n instanceof TypeError&&n.message.includes("fetch")?new s(`Network error: Unable to connect to the Parsefy API. ${n.message}`,"NETWORK_ERROR"):n instanceof TypeError?new s(`Type error: ${n.message}. This may indicate an API response format issue.`,"TYPE_ERROR"):new s(`Unexpected error: ${n instanceof Error?n.message:String(n)}`,"UNKNOWN_ERROR")}}async parseErrorResponse(e){try{return await e.json()}catch{try{return {message:await e.text()||e.statusText}}catch{return {message:e.statusText}}}}};
+exports.APIError=p;exports.DEFAULT_CONFIDENCE_THRESHOLD=u;exports.ExtractionError=y;exports.Parsefy=E;exports.ParsefyError=s;exports.ValidationError=a;

package/dist/index.d.cts

@@ -59,19 +59,19 @@ interface FieldConfidence {
  */
 interface ExtractionMetadata {
     /** Time taken to process the document in milliseconds. */
-    processingTimeMs: number;
+    processing_time_ms: number;
     /** Number of input tokens used. */
-    inputTokens: number;
+    input_tokens: number;
     /** Number of output tokens generated. */
-    outputTokens: number;
+    output_tokens: number;
     /** Number of credits consumed (1 credit = 1 page). */
     credits: number;
     /** Whether the fallback model was triggered for higher accuracy. */
-    fallbackTriggered: boolean;
+    fallback_triggered: boolean;
     /** Overall confidence score for the extraction (0.0 to 1.0). */
-    confidenceScore: number;
+    confidence_score: number;
     /** Per-field confidence details with evidence and explanations. */
-    fieldConfidence: FieldConfidence[];
+    field_confidence: FieldConfidence[];
     /** List of issues or warnings encountered during extraction. */
     issues: string[];
 }
@@ -126,7 +126,7 @@ interface ExtractResult<T> {
  * });
  *
  * // Check per-field confidence and evidence
- * metadata.fieldConfidence.forEach((fc) => {
+ * metadata.field_confidence.forEach((fc) => {
  *   console.log(`${fc.field}: ${fc.score} - "${fc.text}"`);
  * });
  * ```
@@ -194,8 +194,8 @@ declare class Parsefy {
      *   console.log(object.invoice_number); // Fully typed!
      *
      *   // Access field-level confidence and evidence
-     *   console.log(`Overall confidence: ${metadata.confidenceScore}`);
-     *   metadata.fieldConfidence.forEach((fc) => {
+     *   console.log(`Overall confidence: ${metadata.confidence_score}`);
+     *   metadata.field_confidence.forEach((fc) => {
      *     console.log(`${fc.field}: ${fc.score} (${fc.reason}) - "${fc.text}"`);
      *   });
      * }

package/dist/index.d.mts

@@ -59,19 +59,19 @@ interface FieldConfidence {
  */
 interface ExtractionMetadata {
     /** Time taken to process the document in milliseconds. */
-    processingTimeMs: number;
+    processing_time_ms: number;
     /** Number of input tokens used. */
-    inputTokens: number;
+    input_tokens: number;
     /** Number of output tokens generated. */
-    outputTokens: number;
+    output_tokens: number;
     /** Number of credits consumed (1 credit = 1 page). */
     credits: number;
     /** Whether the fallback model was triggered for higher accuracy. */
-    fallbackTriggered: boolean;
+    fallback_triggered: boolean;
     /** Overall confidence score for the extraction (0.0 to 1.0). */
-    confidenceScore: number;
+    confidence_score: number;
     /** Per-field confidence details with evidence and explanations. */
-    fieldConfidence: FieldConfidence[];
+    field_confidence: FieldConfidence[];
     /** List of issues or warnings encountered during extraction. */
     issues: string[];
 }
@@ -126,7 +126,7 @@ interface ExtractResult<T> {
  * });
  *
  * // Check per-field confidence and evidence
- * metadata.fieldConfidence.forEach((fc) => {
+ * metadata.field_confidence.forEach((fc) => {
  *   console.log(`${fc.field}: ${fc.score} - "${fc.text}"`);
  * });
  * ```
@@ -194,8 +194,8 @@ declare class Parsefy {
      *   console.log(object.invoice_number); // Fully typed!
      *
      *   // Access field-level confidence and evidence
-     *   console.log(`Overall confidence: ${metadata.confidenceScore}`);
-     *   metadata.fieldConfidence.forEach((fc) => {
+     *   console.log(`Overall confidence: ${metadata.confidence_score}`);
+     *   metadata.field_confidence.forEach((fc) => {
      *     console.log(`${fc.field}: ${fc.score} (${fc.reason}) - "${fc.text}"`);
      *   });
      * }

package/dist/index.d.ts

@@ -59,19 +59,19 @@ interface FieldConfidence {
  */
 interface ExtractionMetadata {
     /** Time taken to process the document in milliseconds. */
-    processingTimeMs: number;
+    processing_time_ms: number;
     /** Number of input tokens used. */
-    inputTokens: number;
+    input_tokens: number;
     /** Number of output tokens generated. */
-    outputTokens: number;
+    output_tokens: number;
     /** Number of credits consumed (1 credit = 1 page). */
     credits: number;
     /** Whether the fallback model was triggered for higher accuracy. */
-    fallbackTriggered: boolean;
+    fallback_triggered: boolean;
     /** Overall confidence score for the extraction (0.0 to 1.0). */
-    confidenceScore: number;
+    confidence_score: number;
     /** Per-field confidence details with evidence and explanations. */
-    fieldConfidence: FieldConfidence[];
+    field_confidence: FieldConfidence[];
     /** List of issues or warnings encountered during extraction. */
     issues: string[];
 }
@@ -126,7 +126,7 @@ interface ExtractResult<T> {
  * });
  *
  * // Check per-field confidence and evidence
- * metadata.fieldConfidence.forEach((fc) => {
+ * metadata.field_confidence.forEach((fc) => {
  *   console.log(`${fc.field}: ${fc.score} - "${fc.text}"`);
  * });
  * ```
@@ -194,8 +194,8 @@ declare class Parsefy {
      *   console.log(object.invoice_number); // Fully typed!
      *
      *   // Access field-level confidence and evidence
-     *   console.log(`Overall confidence: ${metadata.confidenceScore}`);
-     *   metadata.fieldConfidence.forEach((fc) => {
+     *   console.log(`Overall confidence: ${metadata.confidence_score}`);
+     *   metadata.field_confidence.forEach((fc) => {
      *     console.log(`${fc.field}: ${fc.score} (${fc.reason}) - "${fc.text}"`);
      *   });
      * }

package/dist/index.mjs

@@ -1,2 +1,2 @@
-import {zodToJsonSchema}from'zod-to-json-schema';var u=.85,d={".pdf":"application/pdf",".docx":"application/vnd.openxmlformats-officedocument.wordprocessingml.document"},l=10*1024*1024,h="https://api.parsefy.io",E=6e4;var i=class extends Error{constructor(e,r){super(e),this.name="ParsefyError",this.code=r,typeof Error.captureStackTrace=="function"&&Error.captureStackTrace(this,this.constructor);}},p=class extends i{constructor(e,r,o){super(e),this.name="APIError",this.statusCode=r,this.response=o;}},y=class extends i{constructor(e,r,o){super(e,r),this.name="ExtractionError",this.metadata=o;}},s=class extends i{constructor(e){super(e),this.name="ValidationError";}};function x(){return typeof process<"u"&&process.versions?.node!==void 0}function b(t){let e=zodToJsonSchema(t,{$refStrategy:"none",target:"jsonSchema7"});return "$schema"in e&&delete e.$schema,e}function R(t){let e=t.toLowerCase().match(/\.[^.]+$/)?.[0];return e&&d[e]||null}function w(t){if(!R(t)){let r=Object.keys(d).join(", ");throw new s(`Unsupported file type. Supported types: ${r}`)}}function f(t){if(t===0)throw new s("File is empty");if(t>l){let e=l/1048576;throw new s(`File size exceeds maximum limit of ${e}MB`)}}function F(t){return {object:t.object,metadata:{processingTimeMs:t.metadata.processing_time_ms,inputTokens:t.metadata.input_tokens,outputTokens:t.metadata.output_tokens,credits:t.metadata.credits,fallbackTriggered:t.metadata.fallback_triggered,confidenceScore:t._meta.confidence_score,fieldConfidence:t._meta.field_confidence.map(e=>({field:e.field,score:e.score,reason:e.reason,page:e.page,text:e.text})),issues:t._meta.issues},error:t.error}}function T(t,e){let r=R(e)||"application/octet-stream",o=t.buffer.slice(t.byteOffset,t.byteOffset+t.byteLength);return typeof File<"u"?new File([o],e,{type:r}):new Blob([o],{type:r})}async function k(t){if(!x())throw new s("File path strings are only supported in Node.js. Use File or Blob in the browser.");let e=await import('fs'),r=await import('path');if(!e.existsSync(t))throw new s(`File not found: ${t}`);let o=r.basename(t);w(o);let a=e.readFileSync(t);return f(a.length),{buffer:a,filename:o}}async function _(t){if(typeof t=="string"){let{buffer:e,filename:r}=await k(t);return T(e,r)}if(Buffer.isBuffer(t))return f(t.length),T(t,"document.pdf");if(t instanceof File)return w(t.name),f(t.size),t;if(t instanceof Blob)return f(t.size),t;throw new s("Invalid file input. Expected File, Blob, Buffer, or file path string.")}function P(t){return new Promise(e=>setTimeout(e,t))}function S(t,e=1e3){let r=e*Math.pow(2,t),o=Math.random()*.1*r;return Math.min(r+o,3e4)}var g=class{constructor(e){this.maxRetries=3;let r={};if(typeof e=="string"?r={apiKey:e}:e&&(r=e),this.apiKey=r.apiKey||this.getEnvApiKey(),!this.apiKey)throw new s("API key is required. Provide it in the constructor or set the PARSEFY_API_KEY environment variable.");this.baseUrl=r.baseUrl||h,this.timeout=r.timeout||E;}getEnvApiKey(){return x()&&process.env.PARSEFY_API_KEY||""}async extract(e){let{file:r,schema:o,confidenceThreshold:a}=e,n=b(o),m=await _(r),c=new FormData;return c.append("file",m),c.append("output_schema",JSON.stringify(n)),c.append("confidence_threshold",String(a??.85)),this.makeRequestWithRetry(c)}async makeRequestWithRetry(e,r=0){try{return await this.makeRequest(e)}catch(o){if(o instanceof p&&o.statusCode===429&&r<this.maxRetries){let a=S(r);return await P(a),this.makeRequestWithRetry(e,r+1)}throw o}}async makeRequest(e){let r=`${this.baseUrl}/v1/extract`,o=new AbortController,a=setTimeout(()=>o.abort(),this.timeout);try{let n=await fetch(r,{method:"POST",headers:{Authorization:`Bearer ${this.apiKey}`},body:e,signal:o.signal});if(clearTimeout(a),!n.ok){let c=await this.parseErrorResponse(n);throw new p(c.message||`API request failed with status ${n.status}`,n.status,c)}let m=await n.json();return F(m)}catch(n){throw clearTimeout(a),n instanceof Error&&n.name==="AbortError"?new i(`Request timed out after ${this.timeout}ms`,"TIMEOUT"):n instanceof i?n:n instanceof TypeError?new i("Network error: Unable to connect to the Parsefy API","NETWORK_ERROR"):new i(`Unexpected error: ${n instanceof Error?n.message:String(n)}`,"UNKNOWN_ERROR")}}async parseErrorResponse(e){try{return await e.json()}catch{try{return {message:await e.text()||e.statusText}}catch{return {message:e.statusText}}}}};
-export{p as APIError,u as DEFAULT_CONFIDENCE_THRESHOLD,y as ExtractionError,g as Parsefy,i as ParsefyError,s as ValidationError};
+import {zodToJsonSchema}from'zod-to-json-schema';var u=.85,d={".pdf":"application/pdf",".docx":"application/vnd.openxmlformats-officedocument.wordprocessingml.document"},l=10*1024*1024,g="https://api.parsefy.io",x=6e4;var s=class extends Error{constructor(e,r){super(e),this.name="ParsefyError",this.code=r,typeof Error.captureStackTrace=="function"&&Error.captureStackTrace(this,this.constructor);}},p=class extends s{constructor(e,r,o){super(e),this.name="APIError",this.statusCode=r,this.response=o;}},y=class extends s{constructor(e,r,o){super(e,r),this.name="ExtractionError",this.metadata=o;}},a=class extends s{constructor(e){super(e),this.name="ValidationError";}};function h(){return typeof process<"u"&&process.versions?.node!==void 0}function T(t){let e=zodToJsonSchema(t,{$refStrategy:"none",target:"jsonSchema7"});return "$schema"in e&&delete e.$schema,e}function b(t){let e=t.toLowerCase().match(/\.[^.]+$/)?.[0];return e&&d[e]||null}function _(t){if(!b(t)){let r=Object.keys(d).join(", ");throw new a(`Unsupported file type. Supported types: ${r}`)}}function m(t){if(t===0)throw new a("File is empty");if(t>l){let e=l/1048576;throw new a(`File size exceeds maximum limit of ${e}MB`)}}function w(t){let e=t._meta||{confidence_score:1,field_confidence:[],issues:[]};return {object:t.object,metadata:{processing_time_ms:t.metadata.processing_time_ms,input_tokens:t.metadata.input_tokens,output_tokens:t.metadata.output_tokens,credits:t.metadata.credits,fallback_triggered:t.metadata.fallback_triggered,confidence_score:e.confidence_score,field_confidence:e.field_confidence.map(r=>({field:r.field,score:r.score,reason:r.reason,page:r.page,text:r.text})),issues:e.issues},error:t.error}}function R(t,e){let r=b(e)||"application/octet-stream",o=t.buffer.slice(t.byteOffset,t.byteOffset+t.byteLength);return typeof File<"u"?new File([o],e,{type:r}):new Blob([o],{type:r})}async function I(t){if(!h())throw new a("File path strings are only supported in Node.js. Use File or Blob in the browser.");let e=await import('fs'),r=await import('path');if(!e.existsSync(t))throw new a(`File not found: ${t}`);let o=r.basename(t);_(o);let c=e.readFileSync(t);return m(c.length),{buffer:c,filename:o}}async function F(t){if(typeof t=="string"){let{buffer:e,filename:r}=await I(t);return R(e,r)}if(Buffer.isBuffer(t))return m(t.length),R(t,"document.pdf");if(t instanceof File)return _(t.name),m(t.size),t;if(t instanceof Blob)return m(t.size),t;throw new a("Invalid file input. Expected File, Blob, Buffer, or file path string.")}function P(t){return new Promise(e=>setTimeout(e,t))}function A(t,e=1e3){let r=e*Math.pow(2,t),o=Math.random()*.1*r;return Math.min(r+o,3e4)}var E=class{constructor(e){this.maxRetries=3;let r={};if(typeof e=="string"?r={apiKey:e}:e&&(r=e),this.apiKey=r.apiKey||this.getEnvApiKey(),!this.apiKey)throw new a("API key is required. Provide it in the constructor or set the PARSEFY_API_KEY environment variable.");this.baseUrl=r.baseUrl||g,this.timeout=r.timeout||x;}getEnvApiKey(){return h()&&process.env.PARSEFY_API_KEY||""}async extract(e){let{file:r,schema:o,confidenceThreshold:c}=e,n=T(o),f=await F(r),i=new FormData;return i.append("file",f),i.append("output_schema",JSON.stringify(n)),i.append("confidence_threshold",String(c??.85)),this.makeRequestWithRetry(i)}async makeRequestWithRetry(e,r=0){try{return await this.makeRequest(e)}catch(o){if(o instanceof p&&o.statusCode===429&&r<this.maxRetries){let c=A(r);return await P(c),this.makeRequestWithRetry(e,r+1)}throw o}}async makeRequest(e){let r=`${this.baseUrl}/v1/extract`,o=new AbortController,c=setTimeout(()=>o.abort(),this.timeout);try{let n=await fetch(r,{method:"POST",headers:{Authorization:`Bearer ${this.apiKey}`},body:e,signal:o.signal});if(clearTimeout(c),!n.ok){let i=await this.parseErrorResponse(n);throw new p(i.message||`API request failed with status ${n.status}`,n.status,i)}let f;try{f=await n.json();}catch{throw new s("Failed to parse API response as JSON. The API may have returned an invalid response.","PARSE_ERROR")}try{return w(f)}catch(i){throw new s(`Failed to transform API response: ${i instanceof Error?i.message:String(i)}`,"TRANSFORM_ERROR")}}catch(n){throw clearTimeout(c),n instanceof Error&&n.name==="AbortError"?new s(`Request timed out after ${this.timeout}ms`,"TIMEOUT"):n instanceof s?n:n instanceof TypeError&&n.message.includes("fetch")?new s(`Network error: Unable to connect to the Parsefy API. ${n.message}`,"NETWORK_ERROR"):n instanceof TypeError?new s(`Type error: ${n.message}. This may indicate an API response format issue.`,"TYPE_ERROR"):new s(`Unexpected error: ${n instanceof Error?n.message:String(n)}`,"UNKNOWN_ERROR")}}async parseErrorResponse(e){try{return await e.json()}catch{try{return {message:await e.text()||e.statusText}}catch{return {message:e.statusText}}}}};
+export{p as APIError,u as DEFAULT_CONFIDENCE_THRESHOLD,y as ExtractionError,E as Parsefy,s as ParsefyError,a as ValidationError};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "parsefy",
-  "version": "1.0.2",
+  "version": "1.0.4",
   "description": "Official TypeScript SDK for Parsefy - Financial Document Infrastructure for Developers",
   "author": "",
   "license": "MIT",