app.py

import gradio as gr

from transformers import pipeline

from langchain import PromptTemplate
from langchain.document_loaders import WebBaseLoader
from langchain.schema import StrOutputParser
from langchain.schema.prompt_template import format_document
from langchain.chains import LLMChain
from langchain_google_genai import ChatGoogleGenerativeAI

import ast
import os
import getpass
import matplotlib.pyplot as plt

api_key = os.environ.get('GOOGLE_API_KEY')
if api_key is None:
    raise ValueError("No API key found. Please set the 'GOOGLE_API_KEY' environment variable.")

os.environ['GOOGLE_API_KEY'] = api_key

# Initialize the OCR pipeline
ocr_pipe = pipeline("image-to-text", model="jinhybr/OCR-Donut-CORD")


# Define the initial prompt for the LLM
initial_prompt = '''
You are an IELTS writing examiner. Below, you will receive a series of details pertaining to a writing task. Please evaluate the writing based on the information provided:

1. The type of test (General or Academic)
2. The task number (Task 1 or Task 2)
3. The specific question for this task
4. The author's written response

Please follow these steps for your evaluation:

1. Provide a score on the IELTS scale (0-9 points).
2. Leave a space.
3. Detail the weaknesses and mistakes in the writing.

Evaluate the writing based on the following four criteria:
- Task Response
- Coherence and Cohesion
- Lexical Resource
- Grammatical Range and Accuracy

Respond in this exact format:

Task type: [task type]

Task number: [task number]

score: [score]

Weaknesses and Mistakes:

[Weaknesses and Mistakes]

Notes:

- The text may have been obtained via OCR, which could result in some errors.
- Disregard any text enclosed in <>. They separate different parts of the text.
- If the Task type is empty, try to identify the task type from the question. If you cannot determine the task type, mention that the task type is unclear.
- If the specific question is not present in the task content, mention that the question does not exist in the task response.


Task type: {task_type}

Task number: {task_number}

Question: {question}

Task content: {content}
'''

# Initialize the LLM
llm_model = ChatGoogleGenerativeAI(model="gemini-pro", temperature=0.7, top_p=0.85)

# Define the prompt template
prompt = PromptTemplate(input_variables=['task_type', 'task_number', 'question', 'content'], template=initial_prompt)

# Define the LLM chain
chain = LLMChain(
    llm=llm_model,
    prompt=prompt,
)

def evaluate(task_type, task_number, question, image):
    # Process the image to extract text
    text_content = ocr_pipe(image)
    content = text_content[0]['generated_text']

    # Run the chain
    result = chain.run({
        'task_type': task_type,
        'task_number': task_number,
        'question': question,
        'content': content
    })

    return result

# Create the Gradio interface
inputs = [
    gr.Dropdown(choices=["Academic", "General"], label="Test Type", value="Academic"),
    gr.Dropdown(choices=["Task 1", "Task 2"], label="Task Number", value="Task 1"),
    gr.Textbox(label="Question", value=""),
    gr.Image(type="pil", label="Upload Image")
]


outputs = gr.Markdown(label="Result")

gr.Interface(fn=evaluate, inputs=inputs, outputs=outputs, title="IELTS Writing Evaluation").launch(share=True)