# set path
import glob, os, sys
sys.path.append('../utils')
#import needed libraries
import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import streamlit as st
from utils.policyaction_classifier import load_policyactionClassifier, policyaction_classification
import logging
logger = logging.getLogger(__name__)
from utils.config import get_classifier_params
from utils.preprocessing import paraLengthCheck
from io import BytesIO
import xlsxwriter
import plotly.express as px
from pandas.api.types import (
is_categorical_dtype,
is_datetime64_any_dtype,
is_numeric_dtype,
is_object_dtype,
is_list_like)
# Declare all the necessary variables
classifier_identifier = 'policyaction'
params = get_classifier_params(classifier_identifier)
@st.cache_data
def to_excel(df):
# df['Target Validation'] = 'No'
# df['Netzero Validation'] = 'No'
# df['GHG Validation'] = 'No'
# df['Adapt-Mitig Validation'] = 'No'
# df['Sector'] = 'No'
len_df = len(df)
output = BytesIO()
writer = pd.ExcelWriter(output, engine='xlsxwriter')
df.to_excel(writer, index=False, sheet_name='rawdata')
if 'target_hits' in st.session_state:
target_hits = st.session_state['target_hits']
if 'keep' in target_hits.columns:
target_hits = target_hits[target_hits.keep == True]
target_hits = target_hits.reset_index(drop=True)
target_hits.drop(columns = ['keep'], inplace=True)
target_hits.to_excel(writer,index=False,sheet_name = 'Target')
else:
target_hits = target_hits.sort_values(by=['Target Score'], ascending=False)
target_hits = target_hits.reset_index(drop=True)
target_hits.to_excel(writer,index=False,sheet_name = 'Target')
else:
target_hits = df[df['Target Label'] == True]
target_hits.drop(columns=['Target Label','Netzero Score','GHG Score','Action Label',
'Action Score','Policies_Plans Label','Indicator Label',
'Policies_Plans Score','Conditional Score'],inplace=True)
target_hits = target_hits.sort_values(by=['Target Score'], ascending=False)
target_hits = target_hits.reset_index(drop=True)
target_hits.to_excel(writer,index=False,sheet_name = 'Target')
if 'action_hits' in st.session_state:
action_hits = st.session_state['action_hits']
if 'keep' in action_hits.columns:
action_hits = action_hits[action_hits.keep == True]
action_hits = action_hits.reset_index(drop=True)
action_hits.drop(columns = ['keep'], inplace=True)
action_hits.to_excel(writer,index=False,sheet_name = 'Action')
else:
action_hits = action_hits.sort_values(by=['Action Score'], ascending=False)
action_hits = action_hits.reset_index(drop=True)
action_hits.to_excel(writer,index=False,sheet_name = 'Action')
else:
action_hits = df[df['Action Label'] == True]
action_hits.drop(columns=['Target Label','Target Score','Netzero Score',
'Netzero Label','GHG Label',
'GHG Score','Action Label','Policies_Plans Label',
'Policies_Plans Score','Conditional Score'],inplace=True)
action_hits = action_hits.sort_values(by=['Action Score'], ascending=False)
action_hits = action_hits.reset_index(drop=True)
action_hits.to_excel(writer,index=False,sheet_name = 'Action')
if 'policy_hits' in st.session_state:
policy_hits = st.session_state['policy_hits']
if 'keep' in policy_hits.columns:
policy_hits = policy_hits[policy_hits.keep == True]
policy_hits = policy_hits.reset_index(drop=True)
policy_hits.drop(columns = ['keep'], inplace=True)
policy_hits.to_excel(writer,index=False,sheet_name = 'Policy')
else:
policy_hits = policy_hits.sort_values(by=['Policies_Plans Score'], ascending=False)
policy_hits = policy_hits.reset_index(drop=True)
policy_hits.to_excel(writer,index=False,sheet_name = 'Policy')
else:
policy_hits = df[df['Action Label'] == True]
policy_hits.drop(columns=['Target Label','Target Score','Netzero Score',
'Netzero Label','GHG Label',
'GHG Score','Action Label','Policies_Plans Label',
'Action Score','Conditional Score'],inplace=True)
policy_hits = policy_hits.sort_values(by=['Policies_Plans Score'], ascending=False)
policy_hits = policy_hits.reset_index(drop=True)
policy_hits.to_excel(writer,index=False,sheet_name = 'Policy')
# hits = hits.drop(columns = ['Target Score','Netzero Score','GHG Score'])
workbook = writer.book
# worksheet = writer.sheets['Sheet1']
# worksheet.data_validation('L2:L{}'.format(len_df),
# {'validate': 'list',
# 'source': ['No', 'Yes', 'Discard']})
# worksheet.data_validation('M2:L{}'.format(len_df),
# {'validate': 'list',
# 'source': ['No', 'Yes', 'Discard']})
# worksheet.data_validation('N2:L{}'.format(len_df),
# {'validate': 'list',
# 'source': ['No', 'Yes', 'Discard']})
# worksheet.data_validation('O2:L{}'.format(len_df),
# {'validate': 'list',
# 'source': ['No', 'Yes', 'Discard']})
# worksheet.data_validation('P2:L{}'.format(len_df),
# {'validate': 'list',
# 'source': ['No', 'Yes', 'Discard']})
writer.save()
processed_data = output.getvalue()
return processed_data
# def to_excel(df, hits):
# len_df = len(df)
# output = BytesIO()
# writer = pd.ExcelWriter(output, engine='xlsxwriter')
# df.to_excel(writer, index=False, sheet_name='rawdata')
# if 'keep' in hits.columns:
# hits = hits[hits.keep == True]
# hits = hits.reset_index(drop=True)
# hits.drop(columns = ['keep'], inplace=True)
# # hits = hits.drop(columns = ['Target Score','Netzero Score','GHG Score'])
# hits.to_excel(writer,index=False,sheet_name = 'Action')
# workbook = writer.book
# # worksheet = writer.sheets['Sheet1']
# # worksheet.data_validation('L2:L{}'.format(len_df),
# # {'validate': 'list',
# # 'source': ['No', 'Yes', 'Discard']})
# # worksheet.data_validation('M2:L{}'.format(len_df),
# # {'validate': 'list',
# # 'source': ['No', 'Yes', 'Discard']})
# # worksheet.data_validation('N2:L{}'.format(len_df),
# # {'validate': 'list',
# # 'source': ['No', 'Yes', 'Discard']})
# # worksheet.data_validation('O2:L{}'.format(len_df),
# # {'validate': 'list',
# # 'source': ['No', 'Yes', 'Discard']})
# # worksheet.data_validation('P2:L{}'.format(len_df),
# # {'validate': 'list',
# # 'source': ['No', 'Yes', 'Discard']})
# writer.save()
# processed_data = output.getvalue()
# return processed_data
def app():
### Main app code ###
with st.container():
if 'key1' in st.session_state:
df = st.session_state.key1
classifier = load_policyactionClassifier(classifier_name=params['model_name'])
st.session_state['{}_classifier'.format(classifier_identifier)] = classifier
if sum(df['Target Label'] == 'TARGET') > 100:
warning_msg = ": This might take sometime, please sit back and relax."
else:
warning_msg = ""
df = policyaction_classification(haystack_doc=df,
threshold= params['threshold'])
st.session_state.key1 = df
def action_display():
if 'key1' in st.session_state:
df = st.session_state.key1
st.caption(""" **{}** is splitted into **{}** paragraphs/text chunks."""\
.format(os.path.basename(st.session_state['filename']),
len(df)))
hits = df[df['Action Label'] == 'Action']
range_val = min(5,len(hits))
if range_val !=0:
count_action = len(hits)
hits.drop(columns=['Target Label','Target Score','Netzero Score',
'Netzero Label','GHG Label',
'GHG Score','Action Label','Policies_Plans Label',
'Policies_Plans Score','Conditional Score'],inplace=True)
hits = hits.sort_values(by=['Action Score'], ascending=False)
hits = hits.reset_index(drop=True)
st.write('----------------')
st.caption("Filter table to select rows to keep for Action category")
filter_dataframe_action(hits)
# filtered_df = filtered_df[filtered_df.keep == True]
# st.write('Explore the data')
# AgGrid(hits)
with st.sidebar:
st.write('-------------')
df_xlsx = to_excel(df)
st.download_button(label='📥 Download Result',
data=df_xlsx ,
file_name= os.path.splitext(os.path.basename(st.session_state['filename']))[0]+'.xlsx')
def filter_dataframe_action(df: pd.DataFrame) -> pd.DataFrame:
"""
Adds a UI on top of a dataframe to let viewers filter columns
Args:
df (pd.DataFrame): Original dataframe
Returns:
pd.DataFrame: Filtered dataframe
"""
modify = st.checkbox("Add filters")
if not modify:
st.session_state['action_hits'] = df
return
df = df.copy()
# Try to convert datetimes into a standard format (datetime, no timezone)
# for col in df.columns:
# if is_object_dtype(df[col]):
# try:
# df[col] = pd.to_datetime(df[col])
# except Exception:
# pass
# if is_datetime64_any_dtype(df[col]):
# df[col] = df[col].dt.tz_localize(None)
modification_container = st.container()
with modification_container:
to_filter_columns = st.multiselect("Filter dataframe on", df.columns)
for column in to_filter_columns:
left, right = st.columns((1, 20))
left.write("↳")
# Treat columns with < 10 unique values as categorical
if is_categorical_dtype(df[column]):
user_cat_input = right.multiselect(
f"Values for {column}",
df[column].unique(),
default=list(df[column].unique()),
)
df = df[df[column].isin(user_cat_input)]
elif is_numeric_dtype(df[column]):
_min = float(df[column].min())
_max = float(df[column].max())
step = (_max - _min) / 100
user_num_input = right.slider(
f"Values for {column}",
_min,
_max,
(_min, _max),
step=step,
)
df = df[df[column].between(*user_num_input)]
elif is_list_like(df[column]):
list_vals = set(x for lst in df[column].tolist() for x in lst)
user_multi_input = right.multiselect(
f"Values for {column}",
list_vals,
default=list_vals,
)
df['check'] = df[column].apply(lambda x: any(i in x for i in user_multi_input))
df = df[df.check == True]
df.drop(columns = ['check'],inplace=True)
# df[df[column].between(*user_num_input)]
# elif is_datetime64_any_dtype(df[column]):
# user_date_input = right.date_input(
# f"Values for {column}",
# value=(
# df[column].min(),
# df[column].max(),
# ),
# )
# if len(user_date_input) == 2:
# user_date_input = tuple(map(pd.to_datetime, user_date_input))
# start_date, end_date = user_date_input
# df = df.loc[df[column].between(start_date, end_date)]
else:
user_text_input = right.text_input(
f"Substring or regex in {column}",
)
if user_text_input:
df = df[df[column].str.contains(user_text_input)]
df['keep'] = True
df = df[['keep','text','Action Score','Conditional Label',
'Sector Label','Adapt-Mitig Label','Indicator Label','page']]
df = st.data_editor(
df,
column_config={
"keep": st.column_config.CheckboxColumn(
help="Select which rows to keep",
default=False,
)
},
disabled=list(set(df.columns) - {'keep'}),
hide_index=True,
)
st.session_state['action_hits'] = df
return
def filter_dataframe_policy(df: pd.DataFrame) -> pd.DataFrame:
"""
Adds a UI on top of a dataframe to let viewers filter columns
Args:
df (pd.DataFrame): Original dataframe
Returns:
pd.DataFrame: Filtered dataframe
"""
modify = st.checkbox("Add filters")
if not modify:
st.session_state['policy_hits'] = df
return
df = df.copy()
# Try to convert datetimes into a standard format (datetime, no timezone)
# for col in df.columns:
# if is_object_dtype(df[col]):
# try:
# df[col] = pd.to_datetime(df[col])
# except Exception:
# pass
# if is_datetime64_any_dtype(df[col]):
# df[col] = df[col].dt.tz_localize(None)
modification_container = st.container()
with modification_container:
to_filter_columns = st.multiselect("Filter dataframe on", df.columns)
for column in to_filter_columns:
left, right = st.columns((1, 20))
left.write("↳")
# Treat columns with < 10 unique values as categorical
if is_categorical_dtype(df[column]):
user_cat_input = right.multiselect(
f"Values for {column}",
df[column].unique(),
default=list(df[column].unique()),
)
df = df[df[column].isin(user_cat_input)]
elif is_numeric_dtype(df[column]):
_min = float(df[column].min())
_max = float(df[column].max())
step = (_max - _min) / 100
user_num_input = right.slider(
f"Values for {column}",
_min,
_max,
(_min, _max),
step=step,
)
df = df[df[column].between(*user_num_input)]
elif is_list_like(df[column]):
list_vals = set(x for lst in df[column].tolist() for x in lst)
user_multi_input = right.multiselect(
f"Values for {column}",
list_vals,
default=list_vals,
)
df['check'] = df[column].apply(lambda x: any(i in x for i in user_multi_input))
df = df[df.check == True]
df.drop(columns = ['check'],inplace=True)
# df[df[column].between(*user_num_input)]
# elif is_datetime64_any_dtype(df[column]):
# user_date_input = right.date_input(
# f"Values for {column}",
# value=(
# df[column].min(),
# df[column].max(),
# ),
# )
# if len(user_date_input) == 2:
# user_date_input = tuple(map(pd.to_datetime, user_date_input))
# start_date, end_date = user_date_input
# df = df.loc[df[column].between(start_date, end_date)]
else:
user_text_input = right.text_input(
f"Substring or regex in {column}",
)
if user_text_input:
df = df[df[column].str.contains(user_text_input)]
df['keep'] = True
df = df[['keep','text','Policies_Plans Score','Conditional Label',
'Sector Label','Adapt-Mitig Label','Indicator Label','page']]
df = st.data_editor(
df,
column_config={
"keep": st.column_config.CheckboxColumn(
help="Select which rows to keep",
default=False,
)
},
disabled=list(set(df.columns) - {'keep'}),
hide_index=True,
)
st.session_state['policy_hits'] = df
return
def policy_display():
if 'key1' in st.session_state:
df = st.session_state.key1
st.caption(""" **{}** is splitted into **{}** paragraphs/text chunks."""\
.format(os.path.basename(st.session_state['filename']),
len(df)))
hits = df[df['Policies_Plans Label'] == 'Policy and Plans']
range_val = min(5,len(hits))
if range_val !=0:
count_action = len(hits)
hits.drop(columns=['Target Label','Target Score','Netzero Score',
'Netzero Label','GHG Label',
'GHG Score','Action Label','Policies_Plans Label',
'Action Score','Conditional Score'],inplace=True)
hits = hits.sort_values(by=['Policies_Plans Score'], ascending=False)
hits = hits.reset_index(drop=True)
st.write('----------------')
st.caption("Filter table to select rows to keep for Policies and Plans category")
filter_dataframe_policy(hits)
# filtered_df = filtered_df[filtered_df.keep == True]
# st.write('Explore the data')
# AgGrid(hits)
with st.sidebar:
st.write('-------------')
df_xlsx = to_excel(df)
st.download_button(label='📥 Download Result',
data=df_xlsx ,
file_name= os.path.splitext(os.path.basename(st.session_state['filename']))[0]+'.xlsx')
#count_netzero = sum(hits['Netzero Label'] == 'NETZERO')
#count_ghg = sum(hits['GHG Label'] == 'GHG')
#count_economy = sum([True if 'Economy-wide' in x else False
# for x in hits['Sector Label']])
# count_df = df['Target Label'].value_counts()
# count_df = count_df.rename('count')
# count_df = count_df.rename_axis('Target Label').reset_index()
# count_df['Label_def'] = count_df['Target Label'].apply(lambda x: _lab_dict[x])
# fig = px.bar(count_df, y="Label_def", x="count", orientation='h', height=200)
# c1, c2 = st.columns([1,1])
# with c1:
# st.write('**Target Paragraphs**: `{}`'.format(count_target))
# st.write('**NetZero Related Paragraphs**: `{}`'.format(count_netzero))
#
# # st.plotly_chart(fig,use_container_width= True)
#
# count_netzero = sum(hits['Netzero Label'] == 'NETZERO')
# count_ghg = sum(hits['GHG Label'] == 'LABEL_2')
# count_economy = sum([True if 'Economy-wide' in x else False
# for x in hits['Sector Label']])
# with c2:
# st.write('**GHG Related Paragraphs**: `{}`'.format(count_ghg))
# st.write('**Economy-wide Related Paragraphs**: `{}`'.format(count_economy))
# st.write('-------------------')
# hits = hits.sort_values(by=['Relevancy'], ascending=False)
# netzerohit = hits[hits['Netzero Label'] == 'NETZERO']
# if not netzerohit.empty:
# netzerohit = netzerohit.sort_values(by = ['Netzero Score'], ascending = False)
# # st.write('-------------------')
# st.markdown("###### Netzero paragraph ######")
# st.write('**Netzero paragraph** `page {}`: {}'.format(netzerohit.iloc[0]['page'],
# netzerohit.iloc[0]['text'].replace("\n", " ")))
# st.write("")
# else:
# st.info("🤔 No Netzero paragraph found")
# st.write("**Result {}** `page {}` (Relevancy Score: {:.2f})'".format(i+1,hits.iloc[i]['page'],hits.iloc[i]['Relevancy'])")
# st.write('-------------------')
# st.write("")
# st.markdown("###### Top few Action Classified paragraph/text results from list of {} classified paragraphs ######".format(count_action))
# st.markdown("""
""", unsafe_allow_html=True)
# range_val = min(5,len(hits))
# for i in range(range_val):
# # the page number reflects the page that contains the main paragraph
# # according to split limit, the overlapping part can be on a separate page
# st.write('**Result {}** : `page {}`, `Sector: {}`,\
# `Indicators: {}`, `Adapt-Mitig :{}`'\
# .format(i+1,
# hits.iloc[i]['page'], hits.iloc[i]['Sector Label'],
# hits.iloc[i]['Indicator Label'],hits.iloc[i]['Adapt-Mitig Label']))
# st.write("\t Text: \t{}".format(hits.iloc[i]['text'].replace("\n", " ")))
# hits = hits.reset_index(drop =True)
# st.write('----------------')
# st.write('Explore the data')
# st.write(hits)
# df.drop(columns = ['Action_check'],inplace=True)
# df_xlsx = to_excel(df)
# with st.sidebar:
# st.write('-------------')
# st.download_button(label='📥 Download Result',
# data=df_xlsx ,
# file_name= os.path.splitext(st.session_state['filename'])[0]+'.xlsx')
# else:
# st.info("🤔 No Actions found")
# def policy_display():
# if 'key1' in st.session_state:
# df = st.session_state.key1
# df['Policy_check'] = df['Policy-Action Label'].apply(lambda x: True if 'Policies & Plans' in x else False)
# hits = df[df['Policy_check'] == True]
# # hits['GHG Label'] = hits['GHG Label'].apply(lambda i: _lab_dict[i])
# range_val = min(5,len(hits))
# if range_val !=0:
# count_policy = len(hits)
# #count_netzero = sum(hits['Netzero Label'] == 'NETZERO')
# #count_ghg = sum(hits['GHG Label'] == 'GHG')
# #count_economy = sum([True if 'Economy-wide' in x else False
# # for x in hits['Sector Label']])
# # count_df = df['Target Label'].value_counts()
# # count_df = count_df.rename('count')
# # count_df = count_df.rename_axis('Target Label').reset_index()
# # count_df['Label_def'] = count_df['Target Label'].apply(lambda x: _lab_dict[x])
# # fig = px.bar(count_df, y="Label_def", x="count", orientation='h', height=200)
# # c1, c2 = st.columns([1,1])
# # with c1:
# # st.write('**Target Paragraphs**: `{}`'.format(count_target))
# # st.write('**NetZero Related Paragraphs**: `{}`'.format(count_netzero))
# #
# # # st.plotly_chart(fig,use_container_width= True)
# #
# # count_netzero = sum(hits['Netzero Label'] == 'NETZERO')
# # count_ghg = sum(hits['GHG Label'] == 'LABEL_2')
# # count_economy = sum([True if 'Economy-wide' in x else False
# # for x in hits['Sector Label']])
# # with c2:
# # st.write('**GHG Related Paragraphs**: `{}`'.format(count_ghg))
# # st.write('**Economy-wide Related Paragraphs**: `{}`'.format(count_economy))
# # st.write('-------------------')
# # hits = hits.sort_values(by=['Relevancy'], ascending=False)
# # netzerohit = hits[hits['Netzero Label'] == 'NETZERO']
# # if not netzerohit.empty:
# # netzerohit = netzerohit.sort_values(by = ['Netzero Score'], ascending = False)
# # # st.write('-------------------')
# # st.markdown("###### Netzero paragraph ######")
# # st.write('**Netzero paragraph** `page {}`: {}'.format(netzerohit.iloc[0]['page'],
# # netzerohit.iloc[0]['text'].replace("\n", " ")))
# # st.write("")
# # else:
# # st.info("🤔 No Netzero paragraph found")
# # st.write("**Result {}** `page {}` (Relevancy Score: {:.2f})'".format(i+1,hits.iloc[i]['page'],hits.iloc[i]['Relevancy'])")
# # st.write('-------------------')
# st.write("")
# st.markdown("###### Top few Policy/Plans Classified paragraph/text results from list of {} classified paragraphs ######".format(count_policy))
# st.markdown("""
""", unsafe_allow_html=True)
# range_val = min(5,len(hits))
# for i in range(range_val):
# # the page number reflects the page that contains the main paragraph
# # according to split limit, the overlapping part can be on a separate page
# st.write('**Result {}** : `page {}`, `Sector: {}`,\
# `Indicators: {}`, `Adapt-Mitig :{}`'\
# .format(i+1,
# hits.iloc[i]['page'], hits.iloc[i]['Sector Label'],
# hits.iloc[i]['Indicator Label'],hits.iloc[i]['Adapt-Mitig Label']))
# st.write("\t Text: \t{}".format(hits.iloc[i]['text'].replace("\n", " ")))
# hits = hits.reset_index(drop =True)
# st.write('----------------')
# st.write('Explore the data')
# st.write(hits)
# df.drop(columns = ['Policy_check'],inplace=True)
# df_xlsx = to_excel(df)
# with st.sidebar:
# st.write('-------------')
# st.download_button(label='📥 Download Result',
# data=df_xlsx ,
# file_name= os.path.splitext(st.session_state['filename'])[0]+'.xlsx')
# else:
# st.info("🤔 No Policy/Plans found")