Spaces:

rajeshradhakrishnan
/

english-malayalam

Sleeping

App Files Files Community

rajeshradhakrishnan commited on Apr 21, 2023

Commit

f3b1912

•

1 Parent(s): 74a5927

English-Malayalam Translate v10

Browse files

Files changed (7) hide show

Dockerfile +12 -4
main.py +36 -0
requirements.txt +6 -7
static/index.html +31 -0
static/script.js +84 -0
static/style.css +105 -0
translate.py +0 -56

Dockerfile CHANGED Viewed

@@ -5,13 +5,21 @@ FROM python:3.9
 WORKDIR /code
-ARG TRANSFORMERS_CACHE=/code/translate_cache
 COPY ./requirements.txt /code/requirements.txt
 RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 COPY . .
-CMD gunicorn --bind 0.0.0.0:7860 translate:app
-#CMD ["uvicorn", "translate:app", "--host", "0.0.0.0", "--port", "7860"]

 WORKDIR /code
 COPY ./requirements.txt /code/requirements.txt
 RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 COPY . .
+RUN useradd -m -u 1000 user
+USER user
+ENV HOME=/home/user \
+	PATH=/home/user/.local/bin:$PATH
+WORKDIR $HOME/app
+COPY --chown=user . $HOME/app
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860"]

main.py ADDED Viewed

	@@ -0,0 +1,36 @@

+from fastapi import FastAPI
+from fastapi.staticfiles import StaticFiles
+from fastapi.responses import FileResponse
+from fastapi.templating import Jinja2Templates
+from transformers import MBartForConditionalGeneration, MBart50TokenizerFast
+model = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-50-one-to-many-mmt")
+tokenizer = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50-one-to-many-mmt", src_lang="en_XX")
+app = FastAPI()
+templates = Jinja2Templates(directory="templates")
+@app.get("/infer_t5")
+def t5(input):
+    model_inputs = tokenizer(input, return_tensors="pt")
+    # translate from English to Malayalam
+    generated_tokens = model.generate(
+        **model_inputs,
+        forced_bos_token_id=tokenizer.lang_code_to_id["ml_IN"]
+    )
+    output = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
+    return {"output":output}
+app.mount("/", StaticFiles(directory="static", html=True), name="static")
+# @app.get("/")
+# def index() -> FileResponse:
+#     return FileResponse(path="/app/static/index.html", media_type="text/html")
+@app.get("/")
+async def index():
+    apikey = {"APIKEY": os.environ.get("API_KEY")}
+    return templates.TemplateResponse("index.html", {"apikey": apikey})

requirements.txt CHANGED Viewed

@@ -1,7 +1,6 @@
-torch
-transformers
-sentencepiece
-flask
-flask_restful
-flask-cors
-gunicorn==20.1.0

+fastapi==0.74.*
+requests==2.27.*
+sentencepiece==0.1.*
+torch==1.11.*
+transformers==4.*
+uvicorn[standard]==0.17.*

static/index.html ADDED Viewed

	@@ -0,0 +1,31 @@

+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="UTF-8">
+    <title>JavaScipt Open Assistant Clone</title>
+    <link rel="stylesheet" href="style.css">
+</head>
+<body>
+    <section class="side-bar">
+        <button>New Chat</button>
+        <div class="history"></div>
+        <nav>
+            <p>Made by Rajesh</p>
+        </nav>
+    </section>
+    <section  class="main">
+        <h1>Rajesh - Open Assistant</h1>
+        <p id="output"></p>
+        <div class="bottom-section">
+            <div class="input-container">
+                <input>
+                <div id="submit">➢</div>
+            </div>
+        </div>
+        <p class="info">Open Assistant - This is the 4th iteration English
+            supervised-fine-tuning (SFT) model of the Open-Assistant project.
+        </p>
+    </section>
+    <script src="script.js"></script>
+</body>
+</html>

static/script.js ADDED Viewed

	@@ -0,0 +1,84 @@

+const API_KEY =  {{ apikey.value }};
+const translateText = async (text) => {
+    const inferResponse = await fetch(`infer_t5?input=${text}`);
+    const inferJson = await inferResponse.json();
+    return inferJson.output;
+  };
+function generatePrompterAssistantText(inputString) {
+    // Split the input string into an array of sentences
+    const sentences = inputString.split('<|endoftext|>');
+    // Initialize arrays for prompter and assistant text
+    let prompterText = [];
+    let assistantText = [];
+    // Loop through each sentence and add it to the prompter or assistant text array
+    for (let i = 0; i < sentences.length; i++) {
+      // Check if the sentence contains the <prompter> tag
+      if (sentences[i].includes('<|prompter|>')) {
+        // Extract the text within the <prompter> tags and add it to the prompter text array
+        const prompterSentence = sentences[i].replace(/<\|prompter\|>/g, '');
+        prompterText.push(prompterSentence);
+      } else if (sentences[i].includes('<|assistant|>')) {
+        const assistantSentence = sentences[i].replace(/<\|assistant\|>/g, '');
+        // Add the sentence to the assistant text array
+        assistantText.push(assistantSentence);
+      }
+    }
+    // Return the prompter and assistant text arrays
+    return [prompterText, assistantText];
+  }
+const submitButton = document.querySelector('#submit')
+const outPutElement = document.querySelector('#output')
+const inputElement = document.querySelector('input')
+const historyElement = document.querySelector('.history')
+const buttonElement = document.querySelector('button')
+function changeInput(value)
+{
+    console.log(value)
+    const inputElement = document.querySelector('input')
+    inputElement.value = value
+}
+async function getMessage(){
+    const options = {
+        method: "POST",
+        headers: {
+            Authorization: `Bearer ${API_KEY}`,
+            "Content-Type": "application/json"
+        },
+        body: JSON.stringify({
+            inputs: "<|prompter|>" + inputElement.value + "<|endoftext|><|assistant|>"
+        })
+    }
+    try{
+        const response = await fetch("https://api-inference.huggingface.co/models/OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5", options);
+        const data = await response.json()
+        console.log(data[0].generated_text)
+        if(inputElement.value && data && data[0].generated_text){
+            const [prompterText, assistantText] = generatePrompterAssistantText(data[0].generated_text);
+            outPutElement.textContent = = await translateText(assistantText);
+            const pElement = document.createElement('p')
+            pElement.textContent = inputElement.value
+            pElement.addEventListener('click', () => changeInput(pElement.textContent))
+            historyElement.append(pElement)
+        }
+    } catch(error) {
+        console.log(error)
+    }
+}
+submitButton.addEventListener('click', getMessage)
+function clearInput(){
+    inputElement.value = ''
+}
+buttonElement.addEventListener('click', clearInput)

static/style.css ADDED Viewed

	@@ -0,0 +1,105 @@

+@import url('https://fonts.googleapis.com/css2?family=Open+Sans:wght@400;500;600;700;800&display=swap');
+* {
+    color: #fff;
+    font-family: 'Open Sans', sans-serif;
+}
+body {
+    margin: 0;
+    padding: 0;
+    background-color: #343541;
+    display: flex;
+}
+h1{
+    font-size: 33px;
+    font-weight: 600;
+    padding: 200px 0;
+}
+.side-bar{
+    background-color: #202123;
+    width: 244px;
+    height: 100vh;
+    display: flex;
+    flex-direction: column;
+    justify-content: space-between;
+}
+.main{
+    display: flex;
+    flex-direction: column;
+    align-items: center;
+    text-align: center;
+    justify-content: space-between;
+    height: 100vh;
+    width: 100%;
+}
+.bottom-section{
+    width: 100%;
+    display: flex;
+    flex-direction: column;
+    justify-content: center;
+    align-items: center;
+}
+.info{
+    color: rgba(255, 255, 255, 0.5);
+    font-size: 11px;
+    padding: 10px;
+}
+input{
+    border: none;
+    background-color: rgba(255, 255, 255, 0.5);
+    width: 100%;
+    font-size: 20px;
+    padding: 12px 15px;
+    border-radius: 5px;
+    box-shadow: rgba(0, 0, 0, 0.5) 0 54px 55px,
+    rgba(0, 0, 0, 0.5) 0 -12 30px,
+    rgba(0, 0, 0, 0.5) 0 4px 6px,
+    rgba(0, 0, 0, 0.5) 0 12px 3px,
+    rgba(0, 0, 0, 0.5) 0 -3px 5px
+}
+input:focus{
+    outline: none;
+}
+.input-container{
+    position: relative;
+    width: 100%;
+    max-width: 650px;
+}
+.input-container #submit{
+    position: absolute;
+    right: 0;
+    bottom: 15px;
+    cursor: pointer;
+}
+button{
+    border: solid 0.5px rgba(255, 255, 255, 0.5);
+    background-color: transparent;
+    border-radius: 5px;
+    padding: 10px;
+    margin: 10px;
+}
+nav{
+    border-top: solid 0.5px rgba(255, 255, 255, 0.5);
+    padding: 10px;
+    margin: 10px;
+}
+.history{
+    padding: 10px;
+    margin: 10px;
+    display: flex;
+    flex-direction: column;
+    height: 100%;
+}
+.history p{
+    cursor: pointer;
+}

translate.py DELETED Viewed

@@ -1,56 +0,0 @@
-import torch
-import os
-from transformers.pipelines import pipeline
-from transformers import MBartForConditionalGeneration, MBart50TokenizerFast
-from flask import Flask, request
-from flask_restful import Api, Resource
-from flask_cors import CORS
-app = Flask(__name__)
-CORS(app)
-cors = CORS(app, resource={
-    r"/*": {
-        "origins": "*"
-    }
-})
-api = Api(app)
-app.config['CORS_HEADERS'] = 'Content-Type'
-class Classifier():
-    def __init__(self, data_en):
-        self.model = MBartForConditionalGeneration.from_pretrained("facebook/mbart-large-50-one-to-many-mmt")
-        self.tokenizer = MBart50TokenizerFast.from_pretrained("facebook/mbart-large-50-one-to-many-mmt", src_lang="en_XX")
-        self.model_inputs = self.tokenizer(data_en, return_tensors="pt")
-        # translate from English to Malayalam
-        self.generated_tokens = self.model.generate(
-            **self.model_inputs,
-            forced_bos_token_id=self.tokenizer.lang_code_to_id["ml_IN"]
-        )
-        self. translate = self.tokenizer.batch_decode(self.generated_tokens, skip_special_tokens=True)
-        self.data_en = data_en
-    def get_translator(self):
-        output = self.translate(self.data_en)
-        return {'output': output}
-class Translate(Resource):
-    def post(self):
-        try:
-            # Decode json object from the request
-            json_object = request.get_json()
-            data_en = json_object["text"]
-            obj = Classifier(data_en)
-        except Exception as e:
-            return {"Message": "Error in creating Translator object" + str(e)}
-        status = obj.get_translator()
-        return status
-api.add_resource(Translate, '/api/translate')
-if __name__ == '__main__':
-    app.run(host='0.0.0.0', port=7860)