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
150
151
152
153
154
155
156
157
158
|
# pylint: disable=line-too-long,useless-suppression
# coding=utf-8
# ------------------------------------
# Copyright (c) Microsoft Corporation.
# Licensed under the MIT License.
# ------------------------------------
"""
FILE: sample_conversation_pii_with_entity_mask_policy.py
DESCRIPTION:
This sample demonstrates how to run a PII detection action over a conversation
using the `EntityMaskPolicyType` in sync mode, which redacts detected PII by
replacing it with an entity category mask such as `[Person]` or `[Person-1]`.
USAGE:
python sample_conversation_pii_with_entity_mask_policy.py
REQUIRED ENV VARS (for AAD / DefaultAzureCredential):
AZURE_CONVERSATIONS_ENDPOINT
AZURE_CLIENT_ID
AZURE_TENANT_ID
AZURE_CLIENT_SECRET
NOTE:
If you want to use AzureKeyCredential instead, set:
- AZURE_CONVERSATIONS_ENDPOINT
- AZURE_CONVERSATIONS_KEY
"""
# [START conversation_pii_with_entity_mask_policy]
import os
import re
from azure.identity import DefaultAzureCredential
from azure.ai.language.conversations import ConversationAnalysisClient
from azure.ai.language.conversations.models import (
MultiLanguageConversationInput,
TextConversation,
TextConversationItem,
ParticipantRole,
AnalyzeConversationOperationInput,
PiiOperationAction,
ConversationPiiActionContent,
EntityMaskTypePolicyType,
ConversationPiiOperationResult,
ConversationError,
)
def sample_conversation_pii_with_entity_mask_policy():
# settings
endpoint = os.environ["AZURE_CONVERSATIONS_ENDPOINT"]
credential = DefaultAzureCredential()
redacted_verified = []
client = ConversationAnalysisClient(endpoint, credential=credential)
# build input
ml_input = MultiLanguageConversationInput(
conversations=[
TextConversation(
id="1",
language="en",
conversation_items=[
TextConversationItem(
id="1",
participant_id="Agent_1",
role=ParticipantRole.AGENT,
text="Can you provide your name?",
),
TextConversationItem(
id="2",
participant_id="Customer_1",
role=ParticipantRole.CUSTOMER,
text="Hi, my name is John Doe.",
),
TextConversationItem(
id="3",
participant_id="Agent_1",
role=ParticipantRole.AGENT,
text="Thank you John, that has been updated in our system.",
),
],
)
]
)
# action with EntityMaskTypePolicyType
redaction_policy = EntityMaskTypePolicyType()
pii_action = PiiOperationAction(
action_content=ConversationPiiActionContent(redaction_policy=redaction_policy),
name="Conversation PII with Entity Mask Policy",
)
operation_input = AnalyzeConversationOperationInput(
conversation_input=ml_input,
actions=[pii_action],
)
# start long-running job
poller = client.begin_analyze_conversation_job(body=operation_input)
print(f"Operation ID: {poller.details.get('operation_id')}")
# wait for result
paged_actions = poller.result()
# final metadata
d = poller.details
print(f"Job ID: {d.get('job_id')}")
print(f"Status: {d.get('status')}")
if d.get("errors"):
print("Errors:")
for err in d["errors"]:
print(f" Code: {err.code} - {err.message}")
# iterate results
for actions_page in paged_actions:
for action_result in actions_page.task_results or []:
if isinstance(action_result, ConversationPiiOperationResult):
for conversation in action_result.results.conversations or []:
for item in conversation.conversation_items or []:
redacted_text = (item.redacted_content.text or "").strip()
if not redacted_text:
continue
if item.entities and redacted_text:
all_ok = True
for entity in item.entities:
original_text = entity.text or ""
# 1) original PII must be removed
if original_text and original_text in redacted_text:
print(
f"WARNING: Expected entity '{original_text}' to be redacted "
f"but found in: {redacted_text}"
)
all_ok = False
# 2) mask should appear like [Person] or [Person-1]
expected_mask_pattern = rf"\[{re.escape(entity.category)}-?\d*\]"
if not re.search(expected_mask_pattern, redacted_text, flags=re.IGNORECASE):
print(
f"WARNING: Expected entity mask similar to "
f"'[{entity.category}]' but got: {redacted_text}"
)
all_ok = False
if all_ok:
redacted_verified.append(redacted_text)
# [END conversation_pii_with_entity_mask_policy]
def main():
sample_conversation_pii_with_entity_mask_policy()
if __name__ == "__main__":
main()
|
