Update app.py

app.py

@@ -1,79 +1,12 @@
 import chromadb
 import requests
 import chromadb.utils.embedding_functions as embedding_functions
-import bs4
-import json
+from ai4bharat.transliteration import XlitEngine
 import gradio as gr
 import os
-from io import BytesIO
-import base64
+e = XlitEngine("gu", beam_width=10)
 embeddingfunc = embedding_functions.HuggingFaceEmbeddingFunction(api_key=os.environ["hf_token"],model_name="sentence-transformers/all-MiniLM-L6-v2")
-# client = chromadb.PersistentClient(path="booksofjainism")
-client = chromadb.HttpClient("https://shethjenil-chromadb-server.hf.space/",port=443)
-elibbookAI = client.get_or_create_collection("jainebooks")
-allbookdata = json.load(open("jainbooks.json","r"))
-allsearch = [i['search'] for i in allbookdata]
-class jainnlp:
-    @classmethod
-    def books(cls)->list[str]:
-      return list(set(elibbookAI.get(include=[ "documents" ])["documents"]))
-    @classmethod
-    def loaddata(cls,search:str,progress = gr.Progress(),lang:str="gu")->None:
-      for bookdata in allbookdata:
-       if bookdata['search'] == search:
-        bookname = bookdata['title_english']
-        id = bookdata['sr_no']
-        pages = int(bookdata["pages"])
-        if id not in cls.books():
-            for page,content in enumerate(["\n".join(i.split("\n")[3:]) for i in [i for i in bs4.BeautifulSoup(requests.get(f'https://jainqq.org/booktext/{bookname.replace(" ","_")}/{id}').content, 'html.parser').find('div').stripped_strings][::2]]):
-                try:
-                    contenteng = requests.post("https://translate-pa.googleapis.com/v1/translateHtml", headers={"Content-Type": "application/json+protobuf","User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36","X-Goog-Api-Key": "AIzaSyATBXajvzQLTDHEQbcpq0Ihe0vWDHmO520"}, json=[[content,lang,"en"],"wt_lib"]).json()[0][0]
-                    elibbookAI.add(embeddings=embeddingfunc(contenteng),metadatas={"bookname":bookdata['search'],"page":page,"bookid":id,"originalcontent":content,"contenteng":contenteng,"contentimg":bs4.BeautifulSoup(requests.get(f"https://jainqq.org/explore/{id}/{page}").content, "html.parser").find("img",class_="img-fluid").get("src")},ids=f"{id}-{page}",documents=id)            
-                    progress(page/pages)
-                except:
-                    pass
-        return "done"
-    @classmethod
-    def qna(thisclass,query:str,booklist:list[str] = None,notbooklist:list[str] = None,limit:int=1,lang:str="gu")->list:
-      if booklist:
-        booklist = {"bookid": {"$in": booklist}}
-      if notbooklist:
-        notbooklist = {"bookid": {"$nin": notbooklist}}
-      return [i["contentimg"] for i in elibbookAI.query(embeddingfunc(requests.get(f"https://translate.googleapis.com/translate_a/single?client=gtx&sl={lang}&tl=en&dt=t&q={query}").json()[0][0][0]),n_results=limit,where=booklist)["metadatas"][0]]
-    @classmethod
-    def reset(cls,demo:str):
-       client.reset()
-       return "done"
-    @classmethod
-    def linkmaker(cls,text_input:str,filename:str="book.jainebookAI")->str:
-      file_obj = BytesIO()
-      file_obj.write(text_input.encode())
-      file_obj.seek(0)
-      return f'<a href="data:application/octet-stream;base64,{base64.b64encode(file_obj.getvalue()).decode()}" download="{filename}">Download File</a>'
-
-    @classmethod
-    def download(cls,sr_id:str,embedding:bool=True):
-      id = sr_id
-      if embedding:
-          value = elibbookAI.get(include=["embeddings","metadatas"],where={"bookid":id})
-          return cls.linkmaker(json.dumps({"embeddingenable":True,"DOCUMENT":id,"embedding":value["embeddings"],"metadatas":value["metadatas"]}))
-      else:
-        return cls.linkmaker(json.dumps({"embeddingenable":False,"DOCUMENT":id,"metadatas":elibbookAI.get(include=["metadatas"],where={"bookid":id})["metadatas"]}))
-    @classmethod
-    def upload(cls,file:gr.File)->str:
-      if file:
-        file = json.loads(BytesIO(file).read().decode())
-        documentid = file["DOCUMENT"]
-        metadata = file["metadatas"]
-        if file["embeddingenable"]:
-          elibbookAI.add(embeddings=file["embedding"],metadatas=file["metadatas"],documents=[documentid for i in range(len(metadata))],ids=[f"{documentid}-{metadataofpage['page']}" for metadataofpage in metadata])
-        else:
-          elibbookAI.add(embeddings=[embeddingfunc(metadataofpage["contenteng"]) for metadataofpage in metadata],metadatas=file["metadatas"],documents=file["DOCUMENT"],ids=[f"{documentid}-{metadataofpage['page']}" for metadataofpage in metadata])
-        return "done"
-upload = gr.Interface(jainnlp.loaddata, gr.Dropdown(allsearch),gr.Textbox())
-chatref = gr.Interface(jainnlp.qna,gr.Textbox(),gr.Gallery())
-task = gr.Interface(jainnlp.reset,gr.Textbox(),gr.Textbox(),submit_btn="Reset")
-downloadfile = gr.Interface(jainnlp.download,[gr.Dropdown(jainnlp.books()),gr.Checkbox(True,label="AI ALGORITHEM")],gr.HTML())
-uploadfile = gr.Interface(jainnlp.upload,gr.File(file_types=[".jainebookAI"],type="binary"),gr.Textbox())
-if __name__ == "__main__":
-    gr.TabbedInterface([upload,chatref,task,downloadfile,uploadfile],["Upload","Chat","RESET","DOWNLOADAIFILE","UPLOADFILE"]).launch()
+elibbookAI = chromadb.HttpClient("https://shethjenil-chromadb-server.hf.space/",port=443).get_or_create_collection("jainebooks")
+def qna(query:str,limit:int=1)->list:
+    return [i["contentimg"] for i in elibbookAI.query(embeddingfunc(requests.get(f"https://translate.googleapis.com/translate_a/single?client=gtx&sl=gu&tl=en&dt=t&q={e.translit_sentence(query)["gu"]}").json()[0][0][0]),n_results=limit)["metadatas"][0]]
+gr.Interface(qna,[gr.Textbox(),gr.Slider(1, 4, value=1, label="Count")],gr.Gallery()).launch()