274 feature azure document intelligence as a segmentation model

apps/web/.env.example

@@ -1,6 +1,8 @@
 VITE_API_URL=
+# VITE_API_KEY=
+
 VITE_KEYCLOAK_URL=
 VITE_KEYCLOAK_REALM=
 VITE_KEYCLOAK_CLIENT_ID=
 VITE_KEYCLOAK_REDIRECT_URI=http://localhost:5173
-VITE_KEYCLOAK_POST_LOGOUT_REDIRECT_URI=http://localhost:5173
+VITE_KEYCLOAK_POST_LOGOUT_REDIRECT_URI=http://localhost:5173

apps/web/src/components/Upload/ConfigControls.tsx

@@ -171,21 +171,23 @@ export function SegmentProcessingControls({
   onChange,
   showOnlyPage = false,
 }: SegmentProcessingControlsProps) {
-  const [selectedType, setSelectedType] =
-    useState<keyof SegmentProcessing>("Text");
-  const [isDropdownOpen, setIsDropdownOpen] = useState(false);
   const segmentTypes = showOnlyPage
     ? (["Page"] as (keyof SegmentProcessing)[])
-    : (Object.keys(value).filter(
-        (key) => key !== "Page"
-      ) as (keyof SegmentProcessing)[]);
+    : (Object.keys(value)
+      .filter((key) => key !== "Page")
+      .sort() as (keyof SegmentProcessing)[]);
+
+  const defaultSegmentType = segmentTypes[0];
+  const [selectedType, setSelectedType] =
+    useState<keyof SegmentProcessing>(defaultSegmentType);
+  const [isDropdownOpen, setIsDropdownOpen] = useState(false);
   const dropdownRef = useRef<HTMLDivElement>(null);
 
   useEffect(() => {
     if (showOnlyPage && selectedType !== "Page") {
       setSelectedType("Page");
     } else if (!showOnlyPage && selectedType === "Page") {
-      setSelectedType("Text"); // or any other default segment type
+      setSelectedType(defaultSegmentType);
     }
   }, [selectedType, showOnlyPage]);
 
@@ -270,9 +272,8 @@ export function SegmentProcessingControls({
             {segmentTypes.map((type) => (
               <button
                 key={type}
-                className={`segment-dropdown-item ${
-                  selectedType === type ? "active" : ""
-                } ${isSegmentModified(type) ? "modified" : ""}`}
+                className={`segment-dropdown-item ${selectedType === type ? "active" : ""
+                  } ${isSegmentModified(type) ? "modified" : ""}`}
                 onClick={() => handleTypeSelect(type)}
                 type="button"
               >

apps/web/src/components/Upload/UploadMain.tsx

@@ -6,6 +6,7 @@ import {
   SegmentationStrategy,
   DEFAULT_UPLOAD_CONFIG,
   DEFAULT_SEGMENT_PROCESSING,
+  Pipeline,
 } from "../../models/taskConfig.model";
 import "./UploadMain.css";
 import Upload from "./Upload";
@@ -17,6 +18,7 @@ import {
 } from "./ConfigControls";
 import { uploadFile } from "../../services/uploadFileApi";
 import { UploadForm } from "../../models/upload.model";
+import { getEnvConfig, WhenEnabled } from "../../config/env.config";
 
 interface UploadMainProps {
   onSubmit: (config: UploadFormData) => void;
@@ -28,6 +30,7 @@ export default function UploadMain({
   isAuthenticated,
   onUploadSuccess,
 }: UploadMainProps) {
+  const { features } = getEnvConfig();
   const [files, setFiles] = useState<File[]>([]);
   const [config, setConfig] = useState<UploadFormData>(DEFAULT_UPLOAD_CONFIG);
   const [isUploading, setIsUploading] = useState(false);
@@ -71,6 +74,7 @@ export default function UploadMain({
           ocr_strategy: config.ocr_strategy,
           segment_processing: getEffectiveSegmentProcessing(config),
           segmentation_strategy: config.segmentation_strategy,
+          pipeline: config.pipeline,
         };
 
         const response = await uploadFile(uploadPayload);
@@ -111,6 +115,44 @@ export default function UploadMain({
           className={`config-section ${!isAuthenticated ? "disabled" : ""}`}
         >
           <div className="config-grid">
+            {features.pipeline && (
+              <ToggleGroup
+                label={
+                  <Flex gap="2" align="center">
+                    <svg
+                      width="20px"
+                      height="20px"
+                      viewBox="0 0 16 16"
+                      xmlns="http://www.w3.org/2000/svg"
+                      fill="none"
+                    >
+                      <path
+                        fill="#FFF"
+                        fill-rule="evenodd"
+                        d="M2.75 2.5A1.75 1.75 0 001 4.25v1C1 6.216 1.784 7 2.75 7h1a1.75 1.75 0 001.732-1.5H6.5a.75.75 0 01.75.75v3.5A2.25 2.25 0 009.5 12h1.018c.121.848.85 1.5 1.732 1.5h1A1.75 1.75 0 0015 11.75v-1A1.75 1.75 0 0013.25 9h-1a1.75 1.75 0 00-1.732 1.5H9.5a.75.75 0 01-.75-.75v-3.5A2.25 2.25 0 006.5 4H5.482A1.75 1.75 0 003.75 2.5h-1zM2.5 4.25A.25.25 0 012.75 4h1a.25.25 0 01.25.25v1a.25.25 0 01-.25.25h-1a.25.25 0 01-.25-.25v-1zm9.75 6.25a.25.25 0 00-.25.25v1c0 .138.112.25.25.25h1a.25.25 0 00.25-.25v-1a.25.25 0 00-.25-.25h-1z"
+                        clip-rule="evenodd"
+                      />
+                    </svg>
+                    <span>Pipeline</span>
+                  </Flex>
+                }
+                value={config.pipeline || "Default"}
+                onChange={(value) =>
+                  setConfig({
+                    ...config,
+                    pipeline: (features.pipeline
+                      ? value === "Default"
+                        ? undefined
+                        : (value as Pipeline)
+                      : undefined) as WhenEnabled<"pipeline", Pipeline>,
+                  })
+                }
+                options={[
+                  { label: "Default", value: "Default" },
+                  { label: "Azure", value: Pipeline.Azure },
+                ]}
+              />
+            )}
             <ToggleGroup
               label={
                 <Flex gap="2" align="center">

apps/web/src/config/env.config.ts

@@ -0,0 +1,37 @@
+export interface Features {
+  pipeline: boolean;
+  // Add new feature flags here
+  // example: betaFeature: boolean;
+}
+
+export interface EnvConfig {
+  features: Features;
+}
+
+export const getEnvConfig = (): EnvConfig => {
+  return {
+    features: {
+      pipeline: import.meta.env.VITE_FEATURE_FLAG_PIPELINE === "true",
+      // Add new feature implementations here
+    },
+  };
+};
+
+export function validateEnvConfig(): void {
+  const requiredFlags: Array<keyof Features> = ["pipeline"];
+
+  for (const flag of requiredFlags) {
+    const value = import.meta.env[`VITE_FEATURE_FLAG_${flag.toUpperCase()}`];
+    if (value !== "true" && value !== "false") {
+      throw new Error(
+        `VITE_FEATURE_FLAG_${flag.toUpperCase()} must be either "true" or "false"`,
+      );
+    }
+  }
+}
+
+// Type helper for feature-guarded types
+export type WhenEnabled<
+  Flag extends keyof Features,
+  T,
+> = Features[Flag] extends true ? T | undefined : undefined;

apps/web/src/models/taskConfig.model.ts

@@ -133,6 +133,8 @@ export interface JsonSchema {
   schema_type?: string;
 }
 
+import { WhenEnabled } from "../config/env.config";
+
 export interface UploadFormData {
   /** Optional chunk processing configuration */
   chunk_processing?: ChunkProcessing;
@@ -178,6 +180,13 @@ export interface UploadFormData {
    * @default 512
    */
   target_chunk_length?: number;
+
+  /** Pipeline to run after processing */
+  pipeline?: WhenEnabled<"pipeline", Pipeline>;
+}
+
+export enum Pipeline {
+  Azure = "Azure",
 }
 
 const DEFAULT_SEGMENT_CONFIG: SegmentProcessingConfig = {
@@ -200,8 +209,8 @@ const DEFAULT_FORMULA_CONFIG: SegmentProcessingConfig = {
 
 const DEFAULT_PICTURE_CONFIG: SegmentProcessingConfig = {
   crop_image: CroppingStrategy.All,
-  html: GenerationStrategy.LLM,
-  markdown: GenerationStrategy.LLM,
+  html: GenerationStrategy.Auto,
+  markdown: GenerationStrategy.Auto,
 };
 
 export const DEFAULT_SEGMENT_PROCESSING: SegmentProcessing = {
@@ -227,4 +236,5 @@ export const DEFAULT_UPLOAD_CONFIG: UploadFormData = {
   segment_processing: DEFAULT_SEGMENT_PROCESSING,
   json_schema: undefined, // or some default schema if needed
   file: new File([], ""),
+  pipeline: undefined as WhenEnabled<"pipeline", Pipeline>, // Default pipeline
 };

apps/web/src/models/upload.model.ts

@@ -4,7 +4,9 @@ import {
   OcrStrategy,
   SegmentProcessing,
   SegmentationStrategy,
+  Pipeline,
 } from "./taskConfig.model";
+import { WhenEnabled } from "../config/env.config";
 
 export interface UploadForm {
   /** The file to be uploaded */
@@ -30,4 +32,7 @@ export interface UploadForm {
 
   /** Strategy for document segmentation */
   segmentation_strategy?: SegmentationStrategy;
+
+  /** Pipeline to run after processing */
+  pipeline?: WhenEnabled<"pipeline", Pipeline>;
 }

clients/node-client/jest.load.config.js

@@ -0,0 +1,6 @@
+module.exports = {
+  preset: "ts-jest",
+  testEnvironment: "node",
+  testMatch: ["**/__tests__/**/*.load.test.ts"],
+  testTimeout: 300000, // 5 minute timeout for load tests
+};

clients/python-client/src/chunkr_ai/api/config.py

@@ -127,6 +127,8 @@ class Model(str, Enum):
     FAST = "Fast"
     HIGH_QUALITY = "HighQuality"
 
+class PipelineType(str, Enum):
+    AZURE = "Azure"
 
 class Configuration(BaseModel):
     chunk_processing: Optional[ChunkProcessing] = Field(default=None)
@@ -139,6 +141,7 @@ class Configuration(BaseModel):
     ocr_strategy: Optional[OcrStrategy] = Field(default=None)
     segment_processing: Optional[SegmentProcessing] = Field(default=None)
     segmentation_strategy: Optional[SegmentationStrategy] = Field(default=None)
+    pipeline: Optional[PipelineType] = Field(default=None)
 
     @model_validator(mode="before")
     def map_deprecated_fields(cls, values: Dict) -> Dict:

clients/python-client/src/chunkr_ai/models.py

@@ -18,6 +18,7 @@
     SegmentType,
     SegmentationStrategy,
     Status,
+    PipelineType,
 )
 
 from .api.task import TaskResponse
@@ -45,4 +46,5 @@
     "Status",
     "TaskResponse",
     "TaskResponseAsync",
+    "PipelineType",
 ]

clients/python-client/tests/test_chunkr.py

@@ -10,6 +10,7 @@
     GenerationConfig,
     JsonSchema,
     OcrStrategy,
+    PipelineType,
     Property,
     SegmentationStrategy,
     SegmentProcessing,
@@ -411,3 +412,17 @@ async def test_update_task_direct(chunkr_client, sample_path):
     assert task.status == "Succeeded"
     assert task.output is not None
     assert task.configuration.segmentation_strategy == SegmentationStrategy.PAGE
+
+
+@pytest.mark.asyncio
+async def test_pipeline_type(chunkr_client, sample_path):
+    client_type, client = chunkr_client
+    response = (
+        await client.upload(sample_path, Configuration(pipeline=PipelineType.AZURE))
+        if client_type == "async"
+        else client.upload(sample_path, Configuration(pipeline=PipelineType.AZURE))
+    )
+
+    assert response.task_id is not None
+    assert response.status == "Succeeded"
+    assert response.output is not None

compose.yaml

@@ -108,7 +108,7 @@ services:
     env_file:
       - .env
     deploy:
-      replicas: 1
+      replicas: 0
     restart: always
 
   task:
@@ -121,7 +121,7 @@ services:
     env_file:
       - .env
     deploy:
-      replicas: 1
+      replicas: 0
     restart: always
 
   web:
@@ -132,7 +132,7 @@ services:
     env_file:
       - .env
     deploy:
-      replicas: 1
+      replicas: 0
     restart: always
   segmentation:
     image: luminainc/segmentation:df6e5375

core/.gitignore

@@ -40,4 +40,6 @@ cmd.txt
 
 pdfium-binaries
 
-.cargo
+.cargo
+
+azure-analysis-response.json

core/Cargo.toml

@@ -4,6 +4,9 @@ version = "0.0.0"
 edition = "2021"
 default-run = "core"
 
+[features]
+azure = []
+
 [dependencies]
 actix-cors = "0.7.0"
 actix-multipart = "0.7.2"

core/migrations/.gitkeep

@@ -0,0 +1,2 @@
+ALTER TABLE TASKS 
+    ADD COLUMN version TEXT;

core/src/configs/azure_config.rs

@@ -0,0 +1,36 @@
+use config::{Config as ConfigTrait, ConfigError};
+use dotenvy::dotenv_override;
+use serde::{Deserialize, Serialize};
+
+#[derive(Debug, Serialize, Deserialize)]
+pub struct Config {
+    #[serde(default = "default_api_version")]
+    pub api_version: String,
+    pub endpoint: String,
+    pub key: String,
+    #[serde(default = "default_model_id")]
+    pub model_id: String,
+}
+
+fn default_api_version() -> String {
+    "2024-11-30".to_string()
+}
+
+fn default_model_id() -> String {
+    "prebuilt-layout".to_string()
+}
+
+impl Config {
+    pub fn from_env() -> Result<Self, ConfigError> {
+        dotenv_override().ok();
+
+        ConfigTrait::builder()
+            .add_source(
+                config::Environment::default()
+                    .prefix("AZURE")
+                    .separator("__"),
+            )
+            .build()?
+            .try_deserialize()
+    }
+}

core/src/configs/mod.rs

@@ -12,3 +12,6 @@ pub mod stripe_config;
 pub mod throttle_config;
 pub mod user_config;
 pub mod worker_config;
+
+#[cfg(feature = "azure")]
+pub mod azure_config;

core/src/lib.rs

@@ -25,7 +25,6 @@ pub mod pipeline;
 pub mod routes;
 pub mod utils;
 
-use configs::worker_config;
 use jobs::init::init_jobs;
 use middleware::auth::AuthMiddlewareFactory;
 use routes::github::get_github_repo_info;