Fix extract-anything example (#60)

Author: antoniosarosi

extract-anything/backend/pyproject.toml

@@ -5,7 +5,7 @@ description = "Add your description here"
 readme = "README.md"
 requires-python = ">=3.10"
 dependencies = [
-    "baml-py==0.80.1",
+    "baml-py==0.215.0",
     "fastapi[standard]>=0.115.11",
     "httpx>=0.28.1",
     "pdf2image>=1.17.0",

extract-anything/backend/server.py

@@ -2,7 +2,7 @@
 import json
 import base64
 from typing import Any, Callable, Optional, TypeVar
-from baml_py import BamlStream, Image
+from baml_py import BamlStream, Image, Pdf
 
 import httpx
 from fastapi import FastAPI, UploadFile, File, Form, HTTPException
@@ -13,9 +13,6 @@
 from fastapi.middleware.cors import CORSMiddleware
 from baml_client.types import Schema
 from baml_py.errors import BamlError
-from pdf2image import convert_from_bytes
-from PIL import Image as PILImage
-import io
 
 app = FastAPI()
 
@@ -90,7 +87,7 @@ async def execute_baml(
     url: str = Form(None),
     baml_code: str = Form(...),
     return_type: str = Form(...),
-):   
+):
     final_content = await read_input_content(file, content, url)
     tb = TypeBuilder()
     try:
@@ -130,16 +127,11 @@ async def stream_baml():
             yield json.dumps({ "error": str(e) }) + "\n\n"
     return StreamingResponse(stream_baml(), media_type="text/event-stream")
 
-def convert_to_base64(img: PILImage):
-    buffered = io.BytesIO()
-    img.save(buffered, format="JPEG")
-    return Image.from_base64(base64=base64.b64encode(buffered.getvalue()).decode("utf-8"), media_type="image/jpeg")
-
 async def read_input_content(
     file: Optional[UploadFile] = None,
     content: Optional[str] = None,
     url: Optional[str] = None
-) -> str | Image | list[Image]:
+) -> str | Image | Pdf:
     """
     Processes the input from one of the following:
     - file: an uploaded file (image, audio, PDF or text)
@@ -156,10 +148,9 @@ async def read_input_content(
             file_content = await file.read()
             return file_content.decode("utf-8")
         elif file.content_type == "application/pdf":
-            # Convert PDF to images
             file_content = await file.read()
-            images = convert_from_bytes(file_content)
-            return [convert_to_base64(img) for img in images]
+            file_content_base64 = base64.b64encode(file_content).decode("utf-8")
+            return Pdf.from_base64(file_content_base64)
         else:
             file_content = await file.read()
             file_content_base64 = base64.b64encode(file_content).decode("utf-8")

extract-anything/backend/uv.lock

@@ -3,7 +3,7 @@ class Response {
   @@dynamic
 }
 
-function ExecuteBAML(content: string | image | audio | image[]) -> Response {
+function ExecuteBAML(content: string | image | audio | pdf) -> Response {
   client "openai/gpt-4o"
   prompt #"
     Extract the data from the given content.
@@ -22,11 +22,11 @@ test vaibhav_resume {
     class Person {
       name string @description("The full name of the individual")
       email string @description("The email address of the individual")
-      
+
       experience Experience[]
       skills string[]
     }
-    
+
     class Experience {
       position string @description("The role held by the individual")
       company string @description("The company where the experience was gained")

extract-anything/baml_src/execute_baml.baml

@@ -15,14 +15,14 @@ class Schema {
   other_code string
 }
 
-function GenerateBAML(content: string | image | audio | image[]) -> Schema @stream.not_null {
-  client CustomSonnet
+function GenerateBAML(content: string | image | audio | pdf) -> Schema @stream.not_null {
+  client CustomGPT4o
   prompt #"
     Generate BAML schema for the given content.
 
     {{ BAMLBackground() }}
 
-    {{ ctx.output_format(prefix="Answer with this format:\n") }}
+    {{ ctx.output_format }}
 
     {{ _.role('user') }}
     {{ content }}
@@ -41,17 +41,17 @@ template_string BAMLBackground() ##"
       // Optional string fields use ?
       // @description is optional, but if you include it, it goes after the field.
       name string? @description("The name of the object")
-      
+
       // Arrays of primitives
       // arrays cannot be optional.
       tags string[]
-      
+
       // Enums must be declared separately and are optional
       status MyEnum?
-      
+
       // Union types
       type "success" | "error"
-      
+
       // Primitive types
       count int
       enabled bool

extract-anything/baml_src/generate_baml.baml

@@ -10,7 +10,7 @@ generator py_target {
 
     // The version of the BAML package you have installed (e.g. same version as your baml-py or @boundaryml/baml).
     // The BAML VSCode extension version should also match this version.
-    version "0.207.1"
+    version "0.215.0"
 
     // Valid values: "sync", "async"
     // This controls what `b.FunctionName()` will be (sync or async).
@@ -26,7 +26,7 @@ generator ts_target {
 
     // The version of the BAML package you have installed (e.g. same version as your baml-py or @boundaryml/baml).
     // The BAML VSCode extension version should also match this version.
-    version "0.207.1"
+    version "0.215.0"
 
     // Valid values: "sync", "async"
     // This controls what `b.FunctionName()` will be (sync or async).

extract-anything/baml_src/generators.baml

@@ -7,7 +7,7 @@ interface ErrorMessageProps {
 export function ErrorMessage({ error }: ErrorMessageProps) {
   if (!error) return null
 
-  return <div className="mt-6 p-4 bg-foreground text-white rounded-md">
+  return <div className="mt-6 p-4 bg-red-500 text-white rounded-md">
     {/* {JSON.stringify(error)} */}
     <pre>
       <AnsiColorText text={error} />