Create conditional_classifier.py
Browse files
utils/conditional_classifier.py
ADDED
|
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from typing import List, Tuple
|
| 2 |
+
from typing_extensions import Literal
|
| 3 |
+
import logging
|
| 4 |
+
import pandas as pd
|
| 5 |
+
from pandas import DataFrame, Series
|
| 6 |
+
from utils.config import getconfig
|
| 7 |
+
from utils.preprocessing import processingpipeline
|
| 8 |
+
import streamlit as st
|
| 9 |
+
from transformers import pipeline
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
@st.cache_resource
|
| 13 |
+
def load_conditionalClassifier(config_file:str = None, classifier_name:str = None):
|
| 14 |
+
"""
|
| 15 |
+
loads the document classifier using haystack, where the name/path of model
|
| 16 |
+
in HF-hub as string is used to fetch the model object.Either configfile or
|
| 17 |
+
model should be passed.
|
| 18 |
+
1. https://docs.haystack.deepset.ai/reference/document-classifier-api
|
| 19 |
+
2. https://docs.haystack.deepset.ai/docs/document_classifier
|
| 20 |
+
Params
|
| 21 |
+
--------
|
| 22 |
+
config_file: config file path from which to read the model name
|
| 23 |
+
classifier_name: if modelname is passed, it takes a priority if not \
|
| 24 |
+
found then will look for configfile, else raise error.
|
| 25 |
+
Return: document classifier model
|
| 26 |
+
"""
|
| 27 |
+
if not classifier_name:
|
| 28 |
+
if not config_file:
|
| 29 |
+
logging.warning("Pass either model name or config file")
|
| 30 |
+
return
|
| 31 |
+
else:
|
| 32 |
+
config = getconfig(config_file)
|
| 33 |
+
classifier_name = config.get('conditional','MODEL')
|
| 34 |
+
|
| 35 |
+
logging.info("Loading conditional classifier")
|
| 36 |
+
doc_classifier = pipeline("text-classification",
|
| 37 |
+
model=classifier_name,
|
| 38 |
+
top_k =1)
|
| 39 |
+
|
| 40 |
+
return doc_classifier
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
@st.cache_data
|
| 44 |
+
def conditional_classification(haystack_doc:pd.DataFrame,
|
| 45 |
+
threshold:float = 0.8,
|
| 46 |
+
classifier_model:pipeline= None
|
| 47 |
+
)->Tuple[DataFrame,Series]:
|
| 48 |
+
"""
|
| 49 |
+
Text-Classification on the list of texts provided. Classifier provides the
|
| 50 |
+
most appropriate label for each text. It informs if paragraph contains any
|
| 51 |
+
netzero information or not.
|
| 52 |
+
Params
|
| 53 |
+
---------
|
| 54 |
+
haystack_doc: List of haystack Documents. The output of Preprocessing Pipeline
|
| 55 |
+
contains the list of paragraphs in different format,here the list of
|
| 56 |
+
Haystack Documents is used.
|
| 57 |
+
threshold: threshold value for the model to keep the results from classifier
|
| 58 |
+
classifiermodel: you can pass the classifier model directly,which takes priority
|
| 59 |
+
however if not then looks for model in streamlit session.
|
| 60 |
+
In case of streamlit avoid passing the model directly.
|
| 61 |
+
Returns
|
| 62 |
+
----------
|
| 63 |
+
df: Dataframe
|
| 64 |
+
"""
|
| 65 |
+
logging.info("Working on Conditionality Identification")
|
| 66 |
+
haystack_doc['Conditional Label'] = 'NA'
|
| 67 |
+
haystack_doc['Conditional Score'] = 0.0
|
| 68 |
+
haystack_doc['cond_check'] = False
|
| 69 |
+
haystack_doc['cond_check'] = haystack_doc.apply(lambda x: True if (
|
| 70 |
+
(x['Target Label'] == 'TARGET') | (x['Action Label'] == 'Action') |
|
| 71 |
+
(x['Policies_Plans Label'] == 'Policies and Plans')) else
|
| 72 |
+
False, axis=1)
|
| 73 |
+
# we apply Netzero to only paragraphs which are classified as 'Target' related
|
| 74 |
+
temp = haystack_doc[haystack_doc['cond_check'] == True]
|
| 75 |
+
temp = temp.reset_index(drop=True)
|
| 76 |
+
df = haystack_doc[haystack_doc['cond_check'] == False]
|
| 77 |
+
df = df.reset_index(drop=True)
|
| 78 |
+
|
| 79 |
+
if not classifier_model:
|
| 80 |
+
classifier_model = st.session_state['conditional_classifier']
|
| 81 |
+
|
| 82 |
+
results = classifier_model(list(temp.text))
|
| 83 |
+
labels_= [(l[0]['label'],l[0]['score']) for l in results]
|
| 84 |
+
temp['Conditional Label'],temp['Conditional Score'] = zip(*labels_)
|
| 85 |
+
# temp[' Label'] = temp['Netzero Label'].apply(lambda x: _lab_dict[x])
|
| 86 |
+
# merging Target with Non Target dataframe
|
| 87 |
+
df = pd.concat([df,temp])
|
| 88 |
+
df = df.drop(columns = ['cond_check'])
|
| 89 |
+
df = df.reset_index(drop =True)
|
| 90 |
+
df.index += 1
|
| 91 |
+
|
| 92 |
+
return df
|