Spaces:
Running
Running
petrsovadina
commited on
Commit
•
fdf8839
1
Parent(s):
e15afd8
Update openai_fake_data_generator.py
Browse files- openai_fake_data_generator.py +16 -33
openai_fake_data_generator.py
CHANGED
@@ -1,8 +1,6 @@
|
|
1 |
from collections import namedtuple
|
2 |
from typing import Optional
|
3 |
-
|
4 |
import openai
|
5 |
-
from openai import OpenAI, AzureOpenAI
|
6 |
import logging
|
7 |
|
8 |
logger = logging.getLogger("presidio-streamlit")
|
@@ -12,50 +10,42 @@ OpenAIParams = namedtuple(
|
|
12 |
["openai_key", "model", "api_base", "deployment_id", "api_version", "api_type"],
|
13 |
)
|
14 |
|
15 |
-
|
16 |
def call_completion_model(
|
17 |
prompt: str,
|
18 |
openai_params: OpenAIParams,
|
19 |
max_tokens: Optional[int] = 256,
|
20 |
) -> str:
|
21 |
-
"""Creates a request for the OpenAI Completion service and returns the response.
|
22 |
-
|
23 |
-
|
24 |
-
:param openai_params: OpenAI parameters for the completion model
|
25 |
-
:param max_tokens: The maximum number of tokens to generate.
|
26 |
-
"""
|
27 |
if openai_params.api_type.lower() == "azure":
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
|
|
|
|
33 |
)
|
34 |
else:
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
max_tokens=max_tokens,
|
41 |
-
)
|
42 |
|
43 |
return response.choices[0].text.strip()
|
44 |
|
45 |
-
|
46 |
def create_prompt(anonymized_text: str) -> str:
|
47 |
"""
|
48 |
Create the prompt with instructions to GPT-3.
|
49 |
-
|
50 |
-
:param anonymized_text: Text with placeholders instead of PII values, e.g. My name is <PERSON>.
|
51 |
"""
|
52 |
-
|
53 |
prompt = f"""
|
54 |
Your role is to create synthetic text based on de-identified text with placeholders instead of Personally Identifiable Information (PII).
|
55 |
Replace the placeholders (e.g. ,<PERSON>, {{DATE}}, {{ip_address}}) with fake values.
|
56 |
|
57 |
Instructions:
|
58 |
-
|
59 |
a. Use completely random numbers, so every digit is drawn between 0 and 9.
|
60 |
b. Use realistic names that come from diverse genders, ethnicities and countries.
|
61 |
c. If there are no placeholders, return the text as is.
|
@@ -63,13 +53,6 @@ def create_prompt(anonymized_text: str) -> str:
|
|
63 |
e. If PII exists in the input, replace it with fake values in the output.
|
64 |
f. Remove whitespace before and after the generated text
|
65 |
|
66 |
-
input: [[TEXT STARTS]] How do I change the limit on my credit card {{credit_card_number}}?[[TEXT ENDS]]
|
67 |
-
output: How do I change the limit on my credit card 2539 3519 2345 1555?
|
68 |
-
input: [[TEXT STARTS]]<PERSON> was the chief science officer at <ORGANIZATION>.[[TEXT ENDS]]
|
69 |
-
output: Katherine Buckjov was the chief science officer at NASA.
|
70 |
-
input: [[TEXT STARTS]]Cameroon lives in <LOCATION>.[[TEXT ENDS]]
|
71 |
-
output: Vladimir lives in Moscow.
|
72 |
-
|
73 |
input: [[TEXT STARTS]]{anonymized_text}[[TEXT ENDS]]
|
74 |
output:"""
|
75 |
return prompt
|
|
|
1 |
from collections import namedtuple
|
2 |
from typing import Optional
|
|
|
3 |
import openai
|
|
|
4 |
import logging
|
5 |
|
6 |
logger = logging.getLogger("presidio-streamlit")
|
|
|
10 |
["openai_key", "model", "api_base", "deployment_id", "api_version", "api_type"],
|
11 |
)
|
12 |
|
|
|
13 |
def call_completion_model(
|
14 |
prompt: str,
|
15 |
openai_params: OpenAIParams,
|
16 |
max_tokens: Optional[int] = 256,
|
17 |
) -> str:
|
18 |
+
"""Creates a request for the OpenAI Completion service and returns the response."""
|
19 |
+
openai.api_key = openai_params.openai_key
|
20 |
+
|
|
|
|
|
|
|
21 |
if openai_params.api_type.lower() == "azure":
|
22 |
+
openai.api_base = openai_params.api_base
|
23 |
+
openai.api_version = openai_params.api_version
|
24 |
+
openai.api_type = "azure"
|
25 |
+
response = openai.Completion.create(
|
26 |
+
engine=openai_params.deployment_id,
|
27 |
+
prompt=prompt,
|
28 |
+
max_tokens=max_tokens,
|
29 |
)
|
30 |
else:
|
31 |
+
response = openai.Completion.create(
|
32 |
+
model=openai_params.model,
|
33 |
+
prompt=prompt,
|
34 |
+
max_tokens=max_tokens,
|
35 |
+
)
|
|
|
|
|
36 |
|
37 |
return response.choices[0].text.strip()
|
38 |
|
|
|
39 |
def create_prompt(anonymized_text: str) -> str:
|
40 |
"""
|
41 |
Create the prompt with instructions to GPT-3.
|
42 |
+
:param anonymized_text: Text with placeholders instead of PII values.
|
|
|
43 |
"""
|
|
|
44 |
prompt = f"""
|
45 |
Your role is to create synthetic text based on de-identified text with placeholders instead of Personally Identifiable Information (PII).
|
46 |
Replace the placeholders (e.g. ,<PERSON>, {{DATE}}, {{ip_address}}) with fake values.
|
47 |
|
48 |
Instructions:
|
|
|
49 |
a. Use completely random numbers, so every digit is drawn between 0 and 9.
|
50 |
b. Use realistic names that come from diverse genders, ethnicities and countries.
|
51 |
c. If there are no placeholders, return the text as is.
|
|
|
53 |
e. If PII exists in the input, replace it with fake values in the output.
|
54 |
f. Remove whitespace before and after the generated text
|
55 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
input: [[TEXT STARTS]]{anonymized_text}[[TEXT ENDS]]
|
57 |
output:"""
|
58 |
return prompt
|