Py.Cafe
Gemini Chat with REST API and History
- anleitung.txt
- app.py
- requirements.txt
app.py
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
import json
import requests
import streamlit as st
import pycafe
st.set_page_config(page_title="Gemini Chat", layout="centered")
st.title("Gemini Chat (py.cafe + REST)")
# --- API Key via py.cafe Vault ---
API_KEY = pycafe.get_secret("GEMINI_API_KEY")
if not API_KEY:
st.error("GEMINI_API_KEY fehlt im py.cafe Tresor.")
st.stop()
MODEL = "gemini-2.5-flash"
URL = f"https://generativelanguage.googleapis.com/v1beta/models/{MODEL}:generateContent"
# --- Session State (Chat History) ---
if "messages" not in st.session_state:
st.session_state.messages = []
if "last_incomplete" not in st.session_state:
st.session_state.last_incomplete = False
def looks_incomplete(text: str) -> bool:
t = (text or "").rstrip()
if len(t) < 40:
return False
if t.endswith((".", "!", "?", "…", "”", '"', "```")):
return False
return True
# --- Sidebar Controls ---
with st.sidebar:
st.header("Einstellungen")
temperature = st.slider("Temperature", 0.0, 1.0, 0.4, 0.05)
max_tokens = st.slider("Max Output Tokens", 128, 10000, 2048, 128)
request_timeout = st.slider("HTTP Timeout (s)", 20, 180, 90, 10)
system_prompt = st.text_area(
"System Prompt",
value="Antworte präzise auf Deutsch. Wenn du unsicher bist, sage das explizit.",
height=120,
)
if st.button("Chat leeren"):
st.session_state.messages = []
st.session_state.last_incomplete = False
st.rerun()
# --- Render Chat ---
for m in st.session_state.messages:
role = "user" if m["role"] == "user" else "assistant"
with st.chat_message(role):
st.markdown(m["text"])
def build_contents():
contents = []
if system_prompt.strip():
contents.append({
"role": "user",
"parts": [{"text": f"[System]\n{system_prompt.strip()}"}],
})
for m in st.session_state.messages:
role = "user" if m["role"] == "user" else "model"
contents.append({
"role": role,
"parts": [{"text": m["text"]}],
})
return contents
def gemini_generate(user_text: str) -> str:
contents = build_contents()
contents.append({"role": "user", "parts": [{"text": user_text}]})
payload = {
"contents": contents,
"generationConfig": {
"temperature": float(temperature),
"maxOutputTokens": int(max_tokens),
},
}
r = requests.post(
URL,
params={"key": API_KEY},
headers={"Content-Type": "application/json"},
data=json.dumps(payload),
timeout=int(request_timeout),
)
if r.status_code != 200:
st.session_state.last_incomplete = False
return f"API-Fehler {r.status_code}: {r.text}"
data = r.json()
try:
answer = data["candidates"][0]["content"]["parts"][0]["text"].strip()
st.session_state.last_incomplete = looks_incomplete(answer)
return answer
except Exception:
st.session_state.last_incomplete = False
return f"Unerwartetes Antwortformat: {data}"
# --- Input ---
prompt = st.chat_input("Nachricht an Gemini…")
if prompt:
st.session_state.messages.append({"role": "user", "text": prompt})
with st.chat_message("user"):
st.markdown(prompt)
with st.chat_message("assistant"):
with st.spinner("Gemini antwortet…"):
answer = gemini_generate(prompt)
st.markdown(answer)
st.session_state.messages.append({"role": "assistant", "text": answer})
# --- Continue UI if truncated ---
if st.session_state.get("last_incomplete"):
st.info("Antwort wirkt abgeschnitten. Klicke „Weiter“, um exakt fortzusetzen.")
if st.button("Weiter"):
st.session_state.messages.append({
"role": "user",
"text": "Fahre exakt dort fort, wo du aufgehört hast. Wiederhole nichts. Setze den Satz fort."
})
st.rerun()