Starting page number handling for split pdf page (#55)

**Only `SplitPdfHook.ts`, `SplitPdfHook.test.ts` and overlay_client.yaml
files were modified by human. Rest of them were auto generated.**

To run integration tests first run `unstructured-api` on port 8000

Author: mpolomdeepsense

.npmignore

@@ -1,5 +1,4 @@
-/*
-/docs/**/*.md
+**/*
 !/**/*.ts
 !/**/*.js
 !/**/*.map

.speakeasy/gen.lock

@@ -1,19 +1,19 @@
 lockVersion: 2.0.0
 id: f42cb8e6-e2ce-4565-b975-5a9f38b94d5a
 management:
-  docChecksum: b3b369c355fbcab2d3248c8dc72afaa8
+  docChecksum: 666d45deb8d9066b8e19e04a305ca734
   docVersion: 0.0.1
-  speakeasyVersion: 1.235.0
-  generationVersion: 2.298.0
-  releaseVersion: 0.12.2
-  configChecksum: 247f78afa5a6eac6ec03b4806f3929bb
+  speakeasyVersion: 1.272.0
+  generationVersion: 2.312.1
+  releaseVersion: 0.13.0
+  configChecksum: 738cdd9408fdcef300989dc711ba3cd0
   repoURL: https://github.com/Unstructured-IO/unstructured-js-client.git
   repoSubDirectory: .
   installationURL: https://github.com/Unstructured-IO/unstructured-js-client
   published: true
 features:
   typescript:
-    core: 3.7.0
+    core: 3.8.0
     examples: 2.81.3
     globalSecurity: 2.82.9
     globalServerURLs: 2.82.4
@@ -28,6 +28,7 @@ generatedFiles:
   - .eslintrc.js
   - .npmignore
   - RUNTIMES.md
+  - jsr.json
   - package.json
   - src/index.ts
   - src/lib/base64.ts

Makefile

@@ -29,11 +29,16 @@ build:
 ## test-unit:					run unit tests
 .PHONY: test-unit
 test-unit:
-	npx jest --config jest.config.js test/unit
+	npx jest --detectOpenHandles --config jest.config.js test/unit
+
+## test-integration:			run integration tests
+.PHONY: test-integration
+test-integration:
+	npx jest --detectOpenHandles --config jest.config.js test/integration
 
 ## test:					run all tests
 .PHONY: test
-test: test-unit
+test: test-unit test-integration
 
 #############
 # Speakeasy #

README.md

@@ -168,15 +168,15 @@ In order to speed up processing of long PDF files, set `splitPdfPage` parameter
 
 Warning: this feature causes the `parent_id` metadata generation in elements to be disabled, as it requires having context of multiple pages.
 
-The amount of parallel requests is controlled by `SplitPdfHook.parallelLimit`. By default it equals to 5. It can't be more than 15, to avoid too high resource usage and costs.
+The amount of parallel requests is controlled by `UNSTRUCTURED_CLIENT_SPLIT_CALL_THREADS` environmental variable. By default it equals to 5. It can't be more than 15, to avoid too high resource usage and costs.
 
 ```typescript
 import { SplitPdfHook } from "unstructured-client/hooks/custom/SplitPdfHook";
 
 ...
 
-// Modify this parameter to change the limit of parallel request
-SplitPdfHook.parallelLimit = 10;
+// Modify this environmental variable to change the limit of parallel requests
+process.env["UNSTRUCTURED_CLIENT_SPLIT_CALL_THREADS"] = "10";
 
 client.general.partition({
     files: {
@@ -219,14 +219,14 @@ Certain SDK methods accept files as part of a multi-part request. It is possible
 import { openAsBlob } from "node:fs";
 import { UnstructuredClient } from "unstructured-client";
 
-async function run() {
-    const sdk = new UnstructuredClient({
-        security: {
-            apiKeyAuth: "YOUR_API_KEY",
-        },
-    });
+const unstructuredClient = new UnstructuredClient({
+    security: {
+        apiKeyAuth: "YOUR_API_KEY",
+    },
+});
 
-    const result = await sdk.general.partition({
+async function run() {
+    const result = await unstructuredClient.general.partition({
         chunkingStrategy: "by_title",
         combineUnderNChars: 500,
         encoding: "utf-8",

USAGE.md

@@ -3,14 +3,14 @@
 import { openAsBlob } from "node:fs";
 import { UnstructuredClient } from "unstructured-client";
 
-async function run() {
-    const sdk = new UnstructuredClient({
-        security: {
-            apiKeyAuth: "YOUR_API_KEY",
-        },
-    });
+const unstructuredClient = new UnstructuredClient({
+    security: {
+        apiKeyAuth: "YOUR_API_KEY",
+    },
+});
 
-    const result = await sdk.general.partition({
+async function run() {
+    const result = await unstructuredClient.general.partition({
         chunkingStrategy: "by_title",
         combineUnderNChars: 500,
         encoding: "utf-8",

docs/sdk/models/shared/partitionparameters.md

@@ -10,7 +10,7 @@ generation:
   auth:
     oAuth2ClientCredentialsEnabled: false
 typescript:
-  version: 0.12.2
+  version: 0.13.0
   additionalDependencies:
     dependencies:
       async: ^3.2.5
@@ -21,8 +21,10 @@ typescript:
       jest: ^29.7.0
       ts-jest: ^29.1.2
     peerDependencies: {}
+  additionalPackageJSON: {}
   author: Unstructured
   clientServerStatusCodesAsErrors: true
+  enumFormat: enum
   flattenGlobalSecurity: false
   imports:
     option: openapi

gen.yaml

@@ -2,4 +2,5 @@
 module.exports = {
   preset: "ts-jest",
   testEnvironment: "node",
+  silent: true,
 };

jest.config.js

@@ -0,0 +1,26 @@
+
+
+{
+  "name": "unstructured-client",
+  "version": "0.13.0",
+  "exports": {
+    ".": "./src/index.ts",    
+    "./sdk/models/errors": "./src/sdk/models/errors/index.ts",    
+    "./sdk/models/shared": "./src/sdk/models/shared/index.ts",    
+    "./sdk/models/operations": "./src/sdk/models/operations/index.ts",
+    "./lib/config": "./src/lib/config.ts",
+    "./lib/http": "./src/lib/http.ts",
+    "./lib/retries": "./src/lib/retries.ts",
+    "./lib/sdks": "./src/lib/sdks.ts",
+    "./types": "./src/sdk/types/index.ts"
+  },
+  "publish": {
+    "include": [
+      "LICENSE",
+      "README.md",
+      "RUNTIMES.md",
+      "USAGE.md",
+      "src/**/*.ts"
+    ]
+  }
+}

jsr.json

@@ -11,3 +11,11 @@ actions:
           "title": "Split Pdf Page",
           "description": "Should the pdf file be split at client. Ignored on backend.",
         }
+  - target: $["components"]["schemas"]["partition_parameters"]["properties"]
+    update:
+      "starting_page_number":
+        {
+          "type": "integer",
+          "title": "Starting Page Number",
+          "description": "The real number of the first PDF page.",
+        }

overlay_client.yaml

@@ -1,7 +1,7 @@
 {
   "name": "unstructured-client",
-  "version": "0.12.2",
-  "author": "Unstructured",
+  "version": "0.13.0",
+  "author": "Unstructured",  
   "main": "./index.js",
   "sideEffects": false,
   "repository": {
@@ -14,7 +14,6 @@
     "build": "tsc",
     "prepare": "npm run build"
   },
-  "keywords": [],
   "peerDependencies": {
     "zod": ">= 3"
   },