File size: 5,885 Bytes
3cb5f2b ce5b5d6 3cb5f2b ce5b5d6 3cb5f2b ce5b5d6 3cb5f2b 3984583 3cb5f2b ce5b5d6 3cb5f2b ce5b5d6 3cb5f2b ce5b5d6 3cb5f2b ce5b5d6 3cb5f2b ce5b5d6 3cb5f2b |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 |
# app.py — FastAPI-only (no WebSockets/SSE)
import os
from typing import List, Optional, Tuple
from fastapi import FastAPI
from fastapi.responses import HTMLResponse, JSONResponse
from pydantic import BaseModel, Field
from openai import OpenAI
# --- Your prompts (edit these files in /prompts) ---
from prompts.initial_prompt import INITIAL_PROMPT
from prompts.main_prompt import MAIN_PROMPT
# Read key from Hugging Face Secret (Settings → Variables and secrets)
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
if not OPENAI_API_KEY:
print("WARNING: OPENAI_API_KEY secret is not set. Set it in the Space settings.")
client = OpenAI(api_key=OPENAI_API_KEY)
# ---------- Models for API I/O ----------
class ChatIn(BaseModel):
message: str = Field("", description="The latest user message.")
# Optional: carry conversation history as a list of [user, assistant] turns
history: Optional[List[Tuple[str, str]]] = Field(
default=None,
description="Optional history as list of (user, assistant) tuples."
)
model: str = Field(default="gpt-4o-mini")
temperature: float = Field(default=0.7)
top_p: float = Field(default=0.95)
max_tokens: int = Field(default=600)
class ChatOut(BaseModel):
reply: str
# ---------- Core non-streaming OpenAI call (server-side) ----------
def generate_reply(
user_message: str,
history: Optional[List[Tuple[str, str]]] = None,
model: str = "gpt-4o-mini",
temperature: float = 0.7,
top_p: float = 0.95,
max_tokens: int = 600,
) -> str:
"""
Builds messages as:
system: MAIN_PROMPT
assistant: INITIAL_PROMPT (so users see your intro)
... then optional history [(user, assistant), ...]
user: latest message
"""
messages = [
{"role": "system", "content": MAIN_PROMPT},
{"role": "assistant", "content": INITIAL_PROMPT},
]
if history:
for u, a in history:
if u:
messages.append({"role": "user", "content": u})
if a:
messages.append({"role": "assistant", "content": a})
messages.append({"role": "user", "content": user_message or ""})
try:
resp = client.chat.completions.create(
model=model,
messages=messages,
temperature=temperature,
top_p=top_p,
max_tokens=max_tokens,
)
return resp.choices[0].message.content
except Exception as e:
# Return the error to the UI
return f"Error: {e}"
# ---------- FastAPI app + routes ----------
app = FastAPI(title="Module Chat — Basic Mode")
@app.get("/health")
async def health():
return {"status": "ok"}
@app.post("/api/chat", response_model=ChatOut)
async def api_chat(payload: ChatIn):
reply = generate_reply(
user_message=payload.message,
history=payload.history,
model=payload.model,
temperature=payload.temperature,
top_p=payload.top_p,
max_tokens=payload.max_tokens,
)
return ChatOut(reply=reply)
# Single-page minimal UI that uses ONLY plain HTTPS (no WebSockets)
INDEX_HTML = f"""
<!doctype html>
<html>
<head>
<meta charset="utf-8"/>
<meta name="viewport" content="width=device-width, initial-scale=1"/>
<title>Module Chat — Basic Mode</title>
<style>
:root {{ color-scheme: light dark; }}
body {{ font-family: system-ui, Arial, sans-serif; max-width: 820px; margin: 24px auto; padding: 0 12px; }}
textarea {{ width: 100%; height: 140px; }}
#reply {{ white-space: pre-wrap; border: 1px solid #ddd; padding: 12px; border-radius: 10px; min-height: 120px; }}
button {{ padding: 10px 16px; border-radius: 10px; cursor: pointer; }}
.tip {{ color: #555; font-size: 0.95rem; }}
.row {{ display: flex; gap: 8px; align-items: center; }}
.row > * {{ flex: 1; }}
.small {{ font-size: 12px; color: #666; }}
</style>
</head>
<body>
<h1>Module Chat — Basic Mode</h1>
<p class="tip">
This version uses only normal HTTPS requests (works even on strict campus Wi-Fi).<br/>
<b>Assistant:</b> {INITIAL_PROMPT}
</p>
<textarea id="msg" placeholder="Type your message..."></textarea>
<div class="row">
<button id="send">Send</button>
<label style="flex:0 0 auto">
<input type="checkbox" id="keepHistory" checked /> keep history
</label>
</div>
<h3>Reply</h3>
<div id="reply"></div>
<p class="small">Tip: Press Ctrl/Cmd+Enter to send.</p>
<script>
const btn = document.getElementById("send");
const msg = document.getElementById("msg");
const out = document.getElementById("reply");
const keep = document.getElementById("keepHistory");
let history = [];
async function sendOnce() {{
out.textContent = "Thinking...";
const payload = {{
message: msg.value || "",
history: keep.checked ? history : null
}};
try {{
const res = await fetch("/api/chat", {{
method: "POST",
headers: {{ "Content-Type": "application/json" }},
body: JSON.stringify(payload)
}});
if (!res.ok) {{
const txt = await res.text();
out.textContent = "Server error: " + res.status + " " + txt;
return;
}}
const data = await res.json();
const reply = data.reply || "(no reply)";
out.textContent = reply;
if (keep.checked) {{
history.push([msg.value || "", reply]);
}}
}} catch (e) {{
out.textContent = "Network error: " + e;
}}
}}
btn.onclick = sendOnce;
msg.addEventListener("keydown", (ev) => {{
if (ev.key === "Enter" && (ev.ctrlKey || ev.metaKey)) {{
ev.preventDefault();
sendOnce();
}}
}});
</script>
</body>
</html>
"""
@app.get("/", response_class=HTMLResponse)
async def root():
return INDEX_HTML |