Upload folder using huggingface_hub

.devcontainer/dc.env

@@ -0,0 +1 @@
+OPENAI_API_KEY=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

.devcontainer/docker-compose.yml

@@ -4,6 +4,8 @@ services:
   solar_eyes:
     shm_size: '2gb'
     init: true
+    env_file:
+      - dc.env
     build:
       context: .
       dockerfile: Dockerfile

.gitignore

@@ -12,4 +12,5 @@ yolo*.pt
 *.jpg
 **/.jpg
 pdf_downloads
+node_modules
 

app.py

@@ -1,4 +1,6 @@
 import os
+import logging
+import json
 
 from fastapi import FastAPI, UploadFile
 from fastapi.responses import FileResponse, HTMLResponse, RedirectResponse
@@ -15,7 +17,7 @@ import uuid
 from functools import partial
 from openai import OpenAI
 
-PROMPT = "You are analyzing the spec sheet of a solar panel. Plese answer the following questions, format them with bullets. \n"
+PROMPT = "You are analyzing the spec sheet of a solar panel. Plese answer the following questions, format them as a JSON dictionary.\n"
 
 # from solareyes.sam import SAM
 
@@ -64,6 +66,34 @@ HTML = """
 #     results: Results = model(image)
 #     return results    
 
+def parse_pdf_text(file):
+    pdf = pdfium.PdfDocument(file)
+    all_text = "PDF Extract Text Contents Below: \n\n"
+    for page in pdf:
+        textpage = page.get_textpage()
+        text_all = textpage.get_text_bounded()    
+        all_text += text_all   
+    
+    #use openai to ask questions about text
+    q1 = "What are module dimensions in L x W x H? Result key should be \"module_dimensions\""
+    q2 = "What is the module weight in kilograms? Result key should be \"module_weight\""
+    q3 = "What are the cable lengths in millimeters? Result key should be \"cable_length\""
+    q4 = "What brand, name, or model are the connectors? Result key should be \"connector\""
+    q5 = "How many pieces per container? Prefer 40' HQ or HC, if not available try 53' Result key should be \"pieces_per_container\""
+    q6 = "What is the model number? Result key should be \"model_number\""
+    question = PROMPT + q1 + "\n" + q2 + "\n" + q3 + "\n" + q4 + "\n" + q5 + "\n" + q6 + "\n" + all_text
+    chat_completion = client.chat.completions.create(
+        messages=[
+            {
+                "role": "user",
+                "content": question,
+            }
+        ],
+        model="gpt-3.5-turbo",
+        response_format={ "type": "json_object"}
+    )
+    return chat_completion.choices[0].message.content
+
 
 def segment_solar_panel(image) -> Results:
     # Perform inference
@@ -205,7 +235,6 @@ with gr.Blocks() as demo:
         image = process_pdf_core(pdf)
         return segment_image(image)
 
-
     with gr.Row():
         img_input = gr.Image(label="Upload Image", height=400)
         img_output = gr.AnnotatedImage(height=400)
@@ -221,16 +250,22 @@ with gr.Blocks() as demo:
     
     with gr.Row():
         pdf_input = gr.File(label="Upload PDF", file_types=['pdf'], height=200)
+        img_output_pdf = gr.AnnotatedImage(height=400)
+    with gr.Row():
+        text_input = gr.Textbox(label="Enter Text", placeholder=PROMPT)
+        text_output = gr.Textbox(label="Output", placeholder="Spec analysis will appear here")
         
     pdf_btn = gr.Button("Identify Solar Panel from PDF")
+    pdf_text_btn = gr.Button("Extract specs from PDF Text")
 
     gr.Examples(
         inputs = pdf_input,
         examples = [os.path.join(pdf_dir, file) for file in os.listdir(pdf_dir)],
     )
-
     section_btn.click(segment_image, [img_input], img_output)
-    pdf_btn.click(process_pdf, [pdf_input], img_output)
+    pdf_btn.click(process_pdf, [pdf_input], img_output_pdf)
+    pdf_text_btn.click(parse_pdf_text, [pdf_input], text_output)
+        
     
 #Accept a PDF file, return a jpeg image
 @app.post("/uploadPdf", response_class=FileResponse)
@@ -243,40 +278,17 @@ def extract_image(uploadFile: UploadFile) -> FileResponse:
     jpeg_image.save(filename)
     return FileResponse(filename)
 
-def parse_pdf_text(file):
-    pdf = pdfium.PdfDocument(file)
-    all_text = "PDF Extract Text Contents Below: \n\n"
-    for page in pdf:
-        textpage = page.get_textpage()
-        text_all = textpage.get_text_range()    
-        all_text += text_all   
-    
-    #use openai to ask questions about text
-    q1 = "What are module dimensions in L x W x H?"
-    q2 = "What is the module weight in kilograms?"
-    q3 = "What are the cable lengthes in millimeters?"
-    q4 = "What brand, name, or model are the connectors?"
-    q5 = "How many pieces per container? Prefer 40' HQ or HC, if not available try 53'"
-    q6 = "What is the model number?"
-    question = PROMPT + q1 + "\n" + q2 + "\n" + q3 + "\n" + q4 + "\n" + q5 + "\n" + q6 + "\n" + all_text
-    chat_completion = client.chat.completions.create(
-        messages=[
-            {
-                "role": "user",
-                "content": question,
-            }
-        ],
-        model="gpt-3.5-turbo",
-    )
-    return chat_completion.choices[0].message.content
 
 #Accept a PDF file, return a text summary
 @app.post("/parsePdf")
 def parse_info(uploadFile: UploadFile):
     file = uploadFile.file.read()
+    logging.info(f"Received file {file}")
     answer = parse_pdf_text(file)
-    return {"answer": answer}
+    logging.info(f"Answer: {answer}")
+    return {"answer": json.loads(answer)}
 
+    
 app = gr.mount_gradio_app(app, demo, path="/")
 
 if __name__ == "__main__":

client/run.js

@@ -66,7 +66,8 @@ class PDFExtractor {
 
 // Example usage wrapped in an async IIFE
 (async () => {
-    const converter = new PDFExtractor('http://localhost:7860');
+    // const converter = new PDFExtractor('http://localhost:7860');
+    const converter = new PDFExtractor();
     const pdfPath = '/workspaces/solar_eyes/pdf_downloads/0a0824f0-a5e7-4643-a834-0cea06f36c49.pdf';
     const outputPath = __dirname + '/outputImage2.jpg';
     await converter.convert(pdfPath, outputPath); // Ensure this is awaited if you need sequential execution

client/run.mjs

@@ -67,7 +67,9 @@ class PDFExtractor {
 }
 
 // Usage example
-const extractor = new PDFExtractor('http://localhost:7860');
+// const extractor = new PDFExtractor('http://localhost:7860');
+
+const extractor = new PDFExtractor();
 const pdfPath = join(__dirname, '../pdf_downloads/0a0824f0-a5e7-4643-a834-0cea06f36c49.pdf');
 const outputPath = join(__dirname, 'outputImage.jpg');
 extractor.extract_panel_image(pdfPath, outputPath);