Spaces:
Sleeping
Sleeping
File size: 5,647 Bytes
5e4554d |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 |
import streamlit as st
import pandas as pd
st.set_page_config(page_title="Turkish Review Analysis - via AG", page_icon='📖')
st.header("📖Review Analysis for Your File - TR")
with st.sidebar:
hf_key = st.text_input("HuggingFace Access Key", key="hf_key", type="password")
MODEL_HOTEL = {
"albert": "anilguven/albert_tr_turkish_hotel_reviews", # Add the emoji for the Meta-Llama model
"distilbert": "anilguven/distilbert_tr_turkish_hotel_reviews",
"bert": "anilguven/bert_tr_turkish_hotel_reviews",
"electra": "anilguven/electra_tr_turkish_hotel_reviews",
}
MODEL_MOVIE = {
"albert": "anilguven/albert_tr_turkish_movie_reviews", # Add the emoji for the Meta-Llama model
"distilbert": "anilguven/distilbert_tr_turkish_movie_reviews",
"bert": "anilguven/bert_tr_turkish_movie_reviews",
"electra": "anilguven/electra_tr_turkish_movie_reviews",
}
MODELS = ["albert","distilbert","bert","electra"]
MODEL_TASK = ["Movie review analysis","Hotel review analysis"]
# Use a pipeline as a high-level helper
from transformers import pipeline
# Create a mapping from formatted model names to their original identifiers
def format_model_name(model_key):
name_parts = model_key
formatted_name = ''.join(name_parts) # Join them into a single string with title case
return formatted_name
formatted_names_to_identifiers = {
format_model_name(key): key for key in MODEL_HOTEL.keys()
}
# Debug to ensure names are formatted correctly
#st.write("Formatted Model Names to Identifiers:", formatted_names_to_identifiers
with st.expander("About this app"):
st.write(f"""
1-Upload your file as txt or csv file. Each file contains one sample in the each row.\n
2-Choose your task (movie or hotel review)
3-Choose your model according to your task analysis (negative or positive).\n
4-And model predict your text files. \n
5-Download your test results.
""")
st.text('')
uploaded_file = st.file_uploader(
"Upload a csv or txt file",
type=["csv", "txt"],
help="Scanned documents are not supported yet!",
)
if not uploaded_file or not hf_key:
st.stop()
@st.cache_data
def convert_df(df):
# IMPORTANT: Cache the conversion to prevent computation on every rerun
return df.to_csv().encode("utf-8")
datas = []
try:
if uploaded_file.name.lower().endswith(".csv"):
text = uploaded_file.read().decode("utf-8", errors="replace")
datas = text.split("\n")
with st.expander("Show Datas"):
st.text(datas)
elif uploaded_file.name.lower().endswith(".txt"):
text = uploaded_file.read().decode("utf-8", errors="replace")
datas = text.split("\n")
with st.expander("Show Datas"):
st.text(datas)
else:
raise NotImplementedError(f"File type {uploaded_file.name.split('.')[-1]} not supported")
except Exception as e:
st.error("Error reading file. Make sure the file is not corrupted or encrypted")
st.stop()
task_name: str = st.selectbox("Task", options=MODEL_TASK)
model_select = ''
if task_name == "Movie review analysis": model_select = MODEL_MOVIE
else: model_select = MODEL_HOTEL
model_name: str = st.selectbox("Model", options=MODELS)
selected_model = model_select[model_name]
if not hf_key:
st.info("Please add your HuggingFace Access Key to continue.")
st.stop()
access_token = hf_key
pipe = pipeline("text-classification", model=selected_model, token=access_token)
#from transformers import AutoTokenizer, AutoModelForSequenceClassification
#tokenizer = AutoTokenizer.from_pretrained(selected_model)
#pipe = AutoModelForSequenceClassification.from_pretrained(pretrained_model_name_or_path=selected_model)
# Display the selected model using the formatted name
model_display_name = selected_model # Already formatted
st.write(f"Model being used: `{model_display_name}`")
results=[]
txt = ''
labels=[]
accuracies=[]
values=[]
if st.button("Submit for File Analysis"):#User Review Button
if not hf_key:
st.info("Please add your HuggingFace Access Key to continue.")
st.stop()
else:
label=''
for data in datas:
result = pipe(data)[0]
if result["label"] == "LABEL_0": label = "Negative"
else: label = "Positive"
results.append(data[:-1] + ", " + label + ", " + str(result["score"]*100) + "\n")
labels.append(label)
accuracies.append(str(result["score"]*100))
values.append(data[:-1])
txt += data[:-1] + ", " + label + ", " + str(result["score"]*100) + "\n"
st.text("All files evaluated. You'll download result file.")
if uploaded_file.name.lower().endswith(".txt"):
with st.expander("Show Results"):
st.write(results)
st.download_button('Download Result File', txt, uploaded_file.name.lower()[:-4] + "_results.txt")
elif uploaded_file.name.lower().endswith(".csv"):
dataframe = pd.DataFrame({ "text": values,"label": labels,"accuracy": accuracies})
with st.expander("Show Results"):
st.write(dataframe)
csv = convert_df(dataframe)
st.download_button(label="Download as CSV",data=csv,file_name=uploaded_file.name.lower()[:-4] + "_results.csv",mime="text/csv")
else:
raise NotImplementedError(f"File type not supported")
# with open(result_file) as f:
# st.download_button('Download Txt file', f)
|