Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import tensorflow.compat.v1 as tf
|
2 |
+
import os
|
3 |
+
import shutil
|
4 |
+
import csv
|
5 |
+
import pandas as pd
|
6 |
+
import numpy as np
|
7 |
+
import IPython
|
8 |
+
import streamlit as st
|
9 |
+
import subprocess
|
10 |
+
from itertools import islice
|
11 |
+
import random
|
12 |
+
from transformers import TapasTokenizer, TapasForQuestionAnswering
|
13 |
+
|
14 |
+
tf.get_logger().setLevel('ERROR')
|
15 |
+
|
16 |
+
model_name = 'google/tapas-base-finetuned-wtq'
|
17 |
+
model = TapasForQuestionAnswering.from_pretrained(model_name, local_files_only=False)
|
18 |
+
tokenizer = TapasTokenizer.from_pretrained(model_name)
|
19 |
+
|
20 |
+
|
21 |
+
st.set_option('deprecation.showfileUploaderEncoding', False)
|
22 |
+
|
23 |
+
st.title('Query your Table')
|
24 |
+
|
25 |
+
st.header('Upload CSV file')
|
26 |
+
|
27 |
+
|
28 |
+
uploaded_file = st.file_uploader("Choose your CSV file",type = 'csv')
|
29 |
+
|
30 |
+
placeholder = st.empty()
|
31 |
+
|
32 |
+
if uploaded_file is not None:
|
33 |
+
data = pd.read_csv(uploaded_file)
|
34 |
+
data.replace(',','', regex=True, inplace=True)
|
35 |
+
if st.checkbox('Want to see the data?'):
|
36 |
+
placeholder.dataframe(data)
|
37 |
+
|
38 |
+
st.header('Enter your queries')
|
39 |
+
|
40 |
+
input_queries = st.text_input('Type your queries separated by comma(,)',value='')
|
41 |
+
input_queries = input_queries.split(',')
|
42 |
+
|
43 |
+
colors1 = ["#"+''.join([random.choice('0123456789ABCDEF') for j in range(6)]) for i in range(len(input_queries))]
|
44 |
+
colors2 = ['background-color:'+str(color)+'; color: black' for color in colors1]
|
45 |
+
|
46 |
+
def styling_specific_cell(x,tags,colors):
|
47 |
+
df_styler = pd.DataFrame('', index=x.index, columns=x.columns)
|
48 |
+
for idx,tag in enumerate(tags):
|
49 |
+
for r,c in tag:
|
50 |
+
df_styler.iloc[r, c] = colors[idx]
|
51 |
+
return df_styler
|
52 |
+
|
53 |
+
if st.button('Predict Answers'):
|
54 |
+
with st.spinner('It will take approx a minute'):
|
55 |
+
data = data.astype(str)
|
56 |
+
inputs = tokenizer(table=table, queries=queries, padding='max_length', return_tensors="pt")
|
57 |
+
outputs = model(**inputs)
|
58 |
+
predicted_answer_coordinates, predicted_aggregation_indices = tokenizer.convert_logits_to_predictions( inputs, outputs.logits.detach(), outputs.logits_aggregation.detach())
|
59 |
+
|
60 |
+
id2aggregation = {0: "NONE", 1: "SUM", 2: "AVERAGE", 3:"COUNT"}
|
61 |
+
aggregation_predictions_string = [id2aggregation[x] for x in predicted_aggregation_indices]
|
62 |
+
|
63 |
+
answers = []
|
64 |
+
|
65 |
+
for coordinates in predicted_answer_coordinates:
|
66 |
+
if len(coordinates) == 1:
|
67 |
+
# only a single cell:
|
68 |
+
answers.append(table.iat[coordinates[0]])
|
69 |
+
else:
|
70 |
+
# multiple cells
|
71 |
+
cell_values = []
|
72 |
+
for coordinate in coordinates:
|
73 |
+
cell_values.append(table.iat[coordinate])
|
74 |
+
answers.append(", ".join(cell_values))
|
75 |
+
|
76 |
+
st.success('Done! Please check below the answers and its cells highlighted in table above')
|
77 |
+
|
78 |
+
placeholder.dataframe(data.style.apply(styling_specific_cell,tags=predicted_answer_coordinates,colors=colors2,axis=None))
|
79 |
+
|
80 |
+
for query, answer, predicted_agg, c in zip(queries, answers, aggregation_predictions_string, colors1):
|
81 |
+
st.write('\n')
|
82 |
+
st.markdown('<font color={} size=4>**{}**</font>'.format(c,query), unsafe_allow_html=True)
|
83 |
+
st.write('\n')
|
84 |
+
|
85 |
+
if predicted_agg == "NONE" or predicted_agg == 'COUNT':
|
86 |
+
st.markdown('**>** '+str(answer))
|
87 |
+
else:
|
88 |
+
if predicted_agg == 'SUM':
|
89 |
+
st.markdown('**>** '+str(sum(answer.split(','))))
|
90 |
+
else:
|
91 |
+
st.markdown('**>** '+str(np.round(np.mean(answer.split(',')),2)))
|