Spaces:
Paused
Paused
:pencil: [Doc] Readme: Thanks, Features, Run API service, API usage
Browse files
README.md
CHANGED
|
@@ -7,36 +7,63 @@ sdk: docker
|
|
| 7 |
app_port: 22222
|
| 8 |
---
|
| 9 |
|
| 10 |
-
|
| 11 |
|
| 12 |
-
|
| 13 |
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
-
|
| 17 |
|
| 18 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
```bash
|
| 21 |
# pipreqs . --force --mode no-pin
|
| 22 |
pip install -r requirements.txt
|
| 23 |
```
|
| 24 |
|
| 25 |
-
|
| 26 |
|
| 27 |
```bash
|
| 28 |
-
|
| 29 |
```
|
| 30 |
|
| 31 |
-
## Run
|
| 32 |
|
| 33 |
-
|
| 34 |
|
| 35 |
```bash
|
| 36 |
-
|
| 37 |
```
|
| 38 |
|
| 39 |
-
Docker run
|
| 40 |
|
| 41 |
```bash
|
| 42 |
# no proxy
|
|
@@ -46,8 +73,95 @@ sudo docker run -p 22222:22222 bing-chat-api:1.0
|
|
| 46 |
sudo docker run -p 22222:22222 --env http_proxy="http://<server>:<port>" bing-chat-api:1.0
|
| 47 |
```
|
| 48 |
|
| 49 |
-
##
|
| 50 |
-
|
| 51 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
|
| 53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
app_port: 22222
|
| 8 |
---
|
| 9 |
|
| 10 |
+
# Bing-Chat-API
|
| 11 |
|
| 12 |
+
Chat with Bing like what you do with OpenAI API.
|
| 13 |
|
| 14 |
+
## Thanks
|
| 15 |
+
- [EdgeGPT](https://github.com/acheong08/EdgeGPT) by [@acheong08](https://github.com/acheong08)
|
| 16 |
+
- [bingo](https://github.com/weaigc/bingo) by [@weaigc](https://github.com/weaigc)
|
| 17 |
+
- [@ninomae](https://github.com/NINOMAE1995)
|
| 18 |
|
| 19 |
+
## Features
|
| 20 |
|
| 21 |
+
✅ Implemented:
|
| 22 |
+
|
| 23 |
+
- Support all conversation styles in New Bing
|
| 24 |
+
- `precise`, `balanced`, `creative`
|
| 25 |
+
- Enable/Disable search
|
| 26 |
+
- Model names suffixed with `offline` are disabling search
|
| 27 |
+
- `precise-offline`, `balanced-offline`, `creative-offline`
|
| 28 |
+
- Support OpenAI API format
|
| 29 |
+
- Can use api endpoint via official `openai-python` package
|
| 30 |
+
- Support stream response
|
| 31 |
+
- Support system prompt
|
| 32 |
+
- This means you could bring Sydney back!
|
| 33 |
+
- Support infinite-round chat
|
| 34 |
+
- As long as not exceeded the token limit (~32k)
|
| 35 |
+
- Support Docker deployment
|
| 36 |
+
|
| 37 |
+
🔨 In progress:
|
| 38 |
+
- [ ] Enhance performance and reduce session create requests
|
| 39 |
+
- [ ] Authentication with API key
|
| 40 |
+
|
| 41 |
+
## Run API service
|
| 42 |
+
|
| 43 |
+
### Run in Command Line
|
| 44 |
+
|
| 45 |
+
**Install dependencies:**
|
| 46 |
|
| 47 |
```bash
|
| 48 |
# pipreqs . --force --mode no-pin
|
| 49 |
pip install -r requirements.txt
|
| 50 |
```
|
| 51 |
|
| 52 |
+
**Run API:**
|
| 53 |
|
| 54 |
```bash
|
| 55 |
+
python -m apis.chat_api
|
| 56 |
```
|
| 57 |
|
| 58 |
+
## Run via Docker
|
| 59 |
|
| 60 |
+
**Docker build:**
|
| 61 |
|
| 62 |
```bash
|
| 63 |
+
sudo docker build -t bing-chat-api:1.0 . --build-arg http_proxy=$http_proxy --build-arg https_proxy=$https_proxy
|
| 64 |
```
|
| 65 |
|
| 66 |
+
**Docker run:**
|
| 67 |
|
| 68 |
```bash
|
| 69 |
# no proxy
|
|
|
|
| 73 |
sudo docker run -p 22222:22222 --env http_proxy="http://<server>:<port>" bing-chat-api:1.0
|
| 74 |
```
|
| 75 |
|
| 76 |
+
## API Usage
|
| 77 |
+
|
| 78 |
+
### Using `openai-python`
|
| 79 |
+
|
| 80 |
+
See: [examples/chat_with_openai.py](https://github.com/Hansimov/bing-chat-api/blob/main/examples/chat_with_openai.py)
|
| 81 |
+
|
| 82 |
+
```py
|
| 83 |
+
from openai import OpenAI
|
| 84 |
+
|
| 85 |
+
# If runnning this service with proxy, you might need to unset `http(s)_proxy`.
|
| 86 |
+
base_url = "http://localhost:22222"
|
| 87 |
+
api_key = "sk-xxxxx"
|
| 88 |
+
client = OpenAI(base_url=base_url, api_key=api_key)
|
| 89 |
+
response = client.chat.completions.create(
|
| 90 |
+
model="precise",
|
| 91 |
+
messages=[
|
| 92 |
+
{
|
| 93 |
+
"role": "user",
|
| 94 |
+
"content": "search california's weather for me",
|
| 95 |
+
}
|
| 96 |
+
],
|
| 97 |
+
stream=True,
|
| 98 |
+
)
|
| 99 |
+
|
| 100 |
+
for chunk in response:
|
| 101 |
+
if chunk.choices[0].delta.content is not None:
|
| 102 |
+
print(chunk.choices[0].delta.content, end="", flush=True)
|
| 103 |
+
elif chunk.choices[0].finish_reason == "stop":
|
| 104 |
+
print()
|
| 105 |
+
else:
|
| 106 |
+
pass
|
| 107 |
+
```
|
| 108 |
|
| 109 |
+
### Using post requests
|
| 110 |
+
|
| 111 |
+
See: [examples/chat_with_post.py](https://github.com/Hansimov/bing-chat-api/blob/main/examples/chat_with_post.py)
|
| 112 |
+
|
| 113 |
+
```py
|
| 114 |
+
import ast
|
| 115 |
+
import httpx
|
| 116 |
+
import json
|
| 117 |
+
import re
|
| 118 |
+
|
| 119 |
+
# If runnning this service with proxy, you might need to unset `http(s)_proxy`.
|
| 120 |
+
chat_api = "http://localhost:22222"
|
| 121 |
+
api_key = "sk-xxxxx"
|
| 122 |
+
requests_headers = {}
|
| 123 |
+
|
| 124 |
+
requests_payload = {
|
| 125 |
+
"model": "precise",
|
| 126 |
+
"messages": [
|
| 127 |
+
{
|
| 128 |
+
"role": "user",
|
| 129 |
+
"content": "search and tell me today's weather of california",
|
| 130 |
+
}
|
| 131 |
+
],
|
| 132 |
+
"stream": True,
|
| 133 |
+
}
|
| 134 |
+
|
| 135 |
+
with httpx.stream(
|
| 136 |
+
"POST",
|
| 137 |
+
chat_api + "/chat/completions",
|
| 138 |
+
headers=requests_headers,
|
| 139 |
+
json=requests_payload,
|
| 140 |
+
timeout=httpx.Timeout(connect=20, read=60, write=20, pool=None),
|
| 141 |
+
) as response:
|
| 142 |
+
response_content = ""
|
| 143 |
+
for line in response.iter_lines():
|
| 144 |
+
remove_patterns = [r"^\s*data:\s*", r"^\s*\[DONE\]\s*"]
|
| 145 |
+
for pattern in remove_patterns:
|
| 146 |
+
line = re.sub(pattern, "", line).strip()
|
| 147 |
+
|
| 148 |
+
if line:
|
| 149 |
+
try:
|
| 150 |
+
line_data = json.loads(line)
|
| 151 |
+
except Exception as e:
|
| 152 |
+
try:
|
| 153 |
+
line_data = ast.literal_eval(line)
|
| 154 |
+
except:
|
| 155 |
+
print(f"Error: {line}")
|
| 156 |
+
raise e
|
| 157 |
+
delta_data = line_data["choices"][0]["delta"]
|
| 158 |
+
finish_reason = line_data["choices"][0]["finish_reason"]
|
| 159 |
+
if "role" in delta_data:
|
| 160 |
+
role = delta_data["role"]
|
| 161 |
+
if "content" in delta_data:
|
| 162 |
+
delta_content = delta_data["content"]
|
| 163 |
+
response_content += delta_content
|
| 164 |
+
print(delta_content, end="", flush=True)
|
| 165 |
+
if finish_reason == "stop":
|
| 166 |
+
print()
|
| 167 |
+
```
|