Update main.py
Browse files
main.py
CHANGED
|
@@ -2,12 +2,14 @@ import os
|
|
| 2 |
from fastapi import FastAPI, HTTPException
|
| 3 |
from pydantic import BaseModel
|
| 4 |
import uvicorn
|
| 5 |
-
import requests
|
|
|
|
|
|
|
| 6 |
|
| 7 |
app = FastAPI()
|
| 8 |
|
| 9 |
# Define the Gemini API endpoint for the primary model
|
| 10 |
-
primary_url = "https://
|
| 11 |
|
| 12 |
# Define the data model for the request body
|
| 13 |
class Item(BaseModel):
|
|
@@ -22,6 +24,13 @@ class Item(BaseModel):
|
|
| 22 |
repetition_penalty: float = 1.0
|
| 23 |
key: str = None
|
| 24 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 25 |
# Function to generate the response JSON
|
| 26 |
def generate_response_json(item, output, tokens, model_name):
|
| 27 |
return {
|
|
@@ -47,18 +56,19 @@ def generate_response_json(item, output, tokens, model_name):
|
|
| 47 |
|
| 48 |
# Function to call the Gemini API
|
| 49 |
def call_gemini_api(url, input_text, generate_kwargs):
|
|
|
|
| 50 |
headers = {
|
| 51 |
-
"Authorization": f"Bearer {
|
| 52 |
"Content-Type": "application/json"
|
| 53 |
}
|
| 54 |
data = {
|
| 55 |
-
"
|
| 56 |
-
"
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
| 61 |
-
|
| 62 |
}
|
| 63 |
try:
|
| 64 |
response = requests.post(url, headers=headers, json=data)
|
|
@@ -79,14 +89,14 @@ async def generate_text(item: Item = None):
|
|
| 79 |
raise HTTPException(status_code=400, detail="Parameter `input` or `system prompt` is required.")
|
| 80 |
|
| 81 |
input_ = ""
|
| 82 |
-
if item.system_prompt
|
| 83 |
input_ = f"<s>[INST] {item.system_prompt} [/INST] {item.system_output}</s>"
|
| 84 |
-
elif item.system_prompt
|
| 85 |
input_ = f"<s>[INST] {item.system_prompt} [/INST]</s>"
|
| 86 |
-
elif item.system_output
|
| 87 |
input_ = f"<s>{item.system_output}</s>"
|
| 88 |
|
| 89 |
-
if item.templates
|
| 90 |
for num, template in enumerate(item.templates, start=1):
|
| 91 |
input_ += f"\n<s>[INST] Beginning of archived conversation {num} [/INST]</s>"
|
| 92 |
for i in range(0, len(template), 2):
|
|
@@ -95,7 +105,7 @@ async def generate_text(item: Item = None):
|
|
| 95 |
input_ += f"\n<s>[INST] End of archived conversation {num} [/INST]</s>"
|
| 96 |
|
| 97 |
input_ += f"\n<s>[INST] Beginning of active conversation [/INST]</s>"
|
| 98 |
-
if item.history
|
| 99 |
for input_text, output_text in item.history:
|
| 100 |
input_ += f"\n<s>[INST] {input_text} [/INST]"
|
| 101 |
input_ += f"\n{output_text}"
|
|
@@ -117,8 +127,8 @@ async def generate_text(item: Item = None):
|
|
| 117 |
|
| 118 |
tokens = 0
|
| 119 |
response_data = call_gemini_api(primary_url, input_, generate_kwargs)
|
| 120 |
-
output = response_data['
|
| 121 |
-
tokens =
|
| 122 |
|
| 123 |
return generate_response_json(item, output, tokens, primary_url)
|
| 124 |
|
|
@@ -128,6 +138,5 @@ async def generate_text(item: Item = None):
|
|
| 128 |
except Exception as e:
|
| 129 |
raise HTTPException(status_code=500, detail=f"An error occurred: {e}")
|
| 130 |
|
| 131 |
-
if "
|
| 132 |
-
|
| 133 |
-
raise HTTPException(status_code=401, detail="Valid key is required.")
|
|
|
|
| 2 |
from fastapi import FastAPI, HTTPException
|
| 3 |
from pydantic import BaseModel
|
| 4 |
import uvicorn
|
| 5 |
+
import requests
|
| 6 |
+
from google.auth.transport.requests import Request
|
| 7 |
+
from google.oauth2 import id_token
|
| 8 |
|
| 9 |
app = FastAPI()
|
| 10 |
|
| 11 |
# Define the Gemini API endpoint for the primary model
|
| 12 |
+
primary_url = "https://us-east4-aiplatform.googleapis.com/v1/projects/gen-lang-client-0770444709/locations/us-east4/publishers/google/models/gemini-1.5-flash-latest:generateContent"
|
| 13 |
|
| 14 |
# Define the data model for the request body
|
| 15 |
class Item(BaseModel):
|
|
|
|
| 24 |
repetition_penalty: float = 1.0
|
| 25 |
key: str = None
|
| 26 |
|
| 27 |
+
# Function to obtain an OAuth 2.0 access token
|
| 28 |
+
def get_access_token():
|
| 29 |
+
auth_req = Request()
|
| 30 |
+
target_audience = primary_url.split("models")[0] # Extract the base URL
|
| 31 |
+
token = id_token.fetch_id_token(auth_req, target_audience)
|
| 32 |
+
return token
|
| 33 |
+
|
| 34 |
# Function to generate the response JSON
|
| 35 |
def generate_response_json(item, output, tokens, model_name):
|
| 36 |
return {
|
|
|
|
| 56 |
|
| 57 |
# Function to call the Gemini API
|
| 58 |
def call_gemini_api(url, input_text, generate_kwargs):
|
| 59 |
+
access_token = get_access_token()
|
| 60 |
headers = {
|
| 61 |
+
"Authorization": f"Bearer {access_token}",
|
| 62 |
"Content-Type": "application/json"
|
| 63 |
}
|
| 64 |
data = {
|
| 65 |
+
"contents": [{"role": "user", "parts": [{"text": input_text}]}],
|
| 66 |
+
"generationConfig": {
|
| 67 |
+
"temperature": generate_kwargs['temperature'],
|
| 68 |
+
"maxOutputTokens": generate_kwargs['max_new_tokens'],
|
| 69 |
+
"topP": generate_kwargs['top_p'],
|
| 70 |
+
"repetitionPenalty": generate_kwargs['repetition_penalty'],
|
| 71 |
+
}
|
| 72 |
}
|
| 73 |
try:
|
| 74 |
response = requests.post(url, headers=headers, json=data)
|
|
|
|
| 89 |
raise HTTPException(status_code=400, detail="Parameter `input` or `system prompt` is required.")
|
| 90 |
|
| 91 |
input_ = ""
|
| 92 |
+
if item.system_prompt is not None and item.system_output is not None:
|
| 93 |
input_ = f"<s>[INST] {item.system_prompt} [/INST] {item.system_output}</s>"
|
| 94 |
+
elif item.system_prompt is not None:
|
| 95 |
input_ = f"<s>[INST] {item.system_prompt} [/INST]</s>"
|
| 96 |
+
elif item.system_output is not None:
|
| 97 |
input_ = f"<s>{item.system_output}</s>"
|
| 98 |
|
| 99 |
+
if item.templates is not None:
|
| 100 |
for num, template in enumerate(item.templates, start=1):
|
| 101 |
input_ += f"\n<s>[INST] Beginning of archived conversation {num} [/INST]</s>"
|
| 102 |
for i in range(0, len(template), 2):
|
|
|
|
| 105 |
input_ += f"\n<s>[INST] End of archived conversation {num} [/INST]</s>"
|
| 106 |
|
| 107 |
input_ += f"\n<s>[INST] Beginning of active conversation [/INST]</s>"
|
| 108 |
+
if item.history is not None:
|
| 109 |
for input_text, output_text in item.history:
|
| 110 |
input_ += f"\n<s>[INST] {input_text} [/INST]"
|
| 111 |
input_ += f"\n{output_text}"
|
|
|
|
| 127 |
|
| 128 |
tokens = 0
|
| 129 |
response_data = call_gemini_api(primary_url, input_, generate_kwargs)
|
| 130 |
+
output = response_data['candidates'][0]['content']['parts'][0]['text']
|
| 131 |
+
tokens = response_data['usageMetadata']['totalTokenCount']
|
| 132 |
|
| 133 |
return generate_response_json(item, output, tokens, primary_url)
|
| 134 |
|
|
|
|
| 138 |
except Exception as e:
|
| 139 |
raise HTTPException(status_code=500, detail=f"An error occurred: {e}")
|
| 140 |
|
| 141 |
+
if __name__ == "__main__":
|
| 142 |
+
uvicorn.run(app, host="0.0.0.0", port=8000)
|
|
|