File size: 10,430 Bytes
d12a6b6
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
cb01b8b
d12a6b6
 
 
 
 
 
 
 
 
 
 
 
 
cfdf66d
 
d12a6b6
 
 
 
 
 
d239855
d12a6b6
 
 
 
 
 
 
 
 
 
 
 
d239855
d12a6b6
 
 
 
 
 
 
 
 
 
 
 
8472bc0
d12a6b6
 
 
 
 
 
 
 
 
 
 
 
d239855
d12a6b6
 
 
 
 
 
d9e170e
 
 
 
cfdf66d
d9e170e
 
 
 
 
 
 
 
d12a6b6
 
 
 
 
 
d239855
d12a6b6
 
 
 
 
 
cfdf66d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0185608
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
105831b
 
 
cfdf66d
0185608
cfdf66d
0185608
 
 
 
 
 
cfdf66d
 
 
 
 
0185608
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
"""
Configuration constants for the Geminicli2api proxy server.
Centralizes all configuration to avoid duplication across modules.
"""
import os

# API Endpoints
CODE_ASSIST_ENDPOINT = "https://cloudcode-pa.googleapis.com"

# Client Configuration
CLI_VERSION = "0.1.5"  # Match current gemini-cli version

# OAuth Configuration
CLIENT_ID = "681255809395-oo8ft2oprdrnp9e3aqf6av3hmdib135j.apps.googleusercontent.com"
CLIENT_SECRET = "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"
SCOPES = [
    "https://www.googleapis.com/auth/cloud-platform",
    "https://www.googleapis.com/auth/userinfo.email",
    "https://www.googleapis.com/auth/userinfo.profile",
]

# File Paths
SCRIPT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
CREDENTIAL_FILE = os.path.join(SCRIPT_DIR, os.getenv("GOOGLE_APPLICATION_CREDENTIALS", "oauth_creds.json"))

# Authentication
GEMINI_AUTH_PASSWORD = os.getenv("GEMINI_AUTH_PASSWORD", "123456")

# Default Safety Settings for Google API
DEFAULT_SAFETY_SETTINGS = [
    {"category": "HARM_CATEGORY_HARASSMENT", "threshold": "BLOCK_NONE"},
    {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "BLOCK_NONE"},
    {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "BLOCK_NONE"},
    {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "BLOCK_NONE"},
    {"category": "HARM_CATEGORY_CIVIC_INTEGRITY", "threshold": "BLOCK_NONE"}
]

# Base Models (without search variants)
BASE_MODELS = [
    {
        "name": "models/gemini-2.5-pro-preview-05-06",
        "version": "001",
        "displayName": "Gemini 2.5 Pro Preview 05-06",
        "description": "Preview version of Gemini 2.5 Pro from May 6th",
        "inputTokenLimit": 1048576,
        "outputTokenLimit": 65535,
        "supportedGenerationMethods": ["generateContent", "streamGenerateContent"],
        "temperature": 1.0,
        "maxTemperature": 2.0,
        "topP": 0.95,
        "topK": 64
    },
    {
        "name": "models/gemini-2.5-pro-preview-06-05",
        "version": "001",
        "displayName": "Gemini 2.5 Pro Preview 06-05",
        "description": "Preview version of Gemini 2.5 Pro from June 5th",
        "inputTokenLimit": 1048576,
        "outputTokenLimit": 65535,
        "supportedGenerationMethods": ["generateContent", "streamGenerateContent"],
        "temperature": 1.0,
        "maxTemperature": 2.0,
        "topP": 0.95,
        "topK": 64
    },
    {
        "name": "models/gemini-2.5-pro",
        "version": "001",
        "displayName": "Gemini 2.5 Pro",
        "description": "Advanced multimodal model with enhanced capabilities",
        "inputTokenLimit": 1048576,
        "outputTokenLimit": 65535,
        "supportedGenerationMethods": ["generateContent", "streamGenerateContent"],
        "temperature": 1.0,
        "maxTemperature": 2.0,
        "topP": 0.95,
        "topK": 64
    },
    {
        "name": "models/gemini-2.5-flash-preview-05-20",
        "version": "001",
        "displayName": "Gemini 2.5 Flash Preview 05-20",
        "description": "Preview version of Gemini 2.5 Flash from May 20th",
        "inputTokenLimit": 1048576,
        "outputTokenLimit": 65535,
        "supportedGenerationMethods": ["generateContent", "streamGenerateContent"],
        "temperature": 1.0,
        "maxTemperature": 2.0,
        "topP": 0.95,
        "topK": 64
    },
    {
        "name": "models/gemini-2.5-flash-preview-04-17",
        "version": "001",
        "displayName": "Gemini 2.5 Flash Preview 04-17",
        "description": "Preview version of Gemini 2.5 Flash from April 17th",
        "inputTokenLimit": 1048576,
        "outputTokenLimit": 65535,
        "supportedGenerationMethods": ["generateContent", "streamGenerateContent"],
        "temperature": 1.0,
        "maxTemperature": 2.0,
        "topP": 0.95,
        "topK": 64
    },
    {
        "name": "models/gemini-2.5-flash",
        "version": "001",
        "displayName": "Gemini 2.5 Flash",
        "description": "Fast and efficient multimodal model with latest improvements",
        "inputTokenLimit": 1048576,
        "outputTokenLimit": 65535,
        "supportedGenerationMethods": ["generateContent", "streamGenerateContent"],
        "temperature": 1.0,
        "maxTemperature": 2.0,
        "topP": 0.95,
        "topK": 64
    }
]

# Generate search variants for applicable models
def _generate_search_variants():
    """Generate search variants for models that support content generation."""
    search_models = []
    for model in BASE_MODELS:
        # Only add search variants for models that support content generation
        if "generateContent" in model["supportedGenerationMethods"]:
            search_variant = model.copy()
            search_variant["name"] = model["name"] + "-search"
            search_variant["displayName"] = model["displayName"] + " with Google Search"
            search_variant["description"] = model["description"] + " (includes Google Search grounding)"
            search_models.append(search_variant)
    return search_models

# Generate thinking variants for applicable models
def _generate_thinking_variants():
    """Generate nothinking and maxthinking variants for models that support thinking."""
    thinking_models = []
    for model in BASE_MODELS:
        # Only add thinking variants for models that support content generation
        # and contain "gemini-2.5-flash" or "gemini-2.5-pro" in their name
        if ("generateContent" in model["supportedGenerationMethods"] and
            ("gemini-2.5-flash" in model["name"] or "gemini-2.5-pro" in model["name"])):
            
            # Add -nothinking variant
            nothinking_variant = model.copy()
            nothinking_variant["name"] = model["name"] + "-nothinking"
            nothinking_variant["displayName"] = model["displayName"] + " (No Thinking)"
            nothinking_variant["description"] = model["description"] + " (thinking disabled)"
            thinking_models.append(nothinking_variant)
            
            # Add -maxthinking variant
            maxthinking_variant = model.copy()
            maxthinking_variant["name"] = model["name"] + "-maxthinking"
            maxthinking_variant["displayName"] = model["displayName"] + " (Max Thinking)"
            maxthinking_variant["description"] = model["description"] + " (maximum thinking budget)"
            thinking_models.append(maxthinking_variant)
    return thinking_models

# Generate combined variants (search + thinking combinations)
def _generate_combined_variants():
    """Generate combined search and thinking variants."""
    combined_models = []
    for model in BASE_MODELS:
        # Only add combined variants for models that support content generation
        # and contain "gemini-2.5-flash" or "gemini-2.5-pro" in their name
        if ("generateContent" in model["supportedGenerationMethods"] and
            ("gemini-2.5-flash" in model["name"] or "gemini-2.5-pro" in model["name"])):
            
            # search + nothinking
            search_nothinking = model.copy()
            search_nothinking["name"] = model["name"] + "-search-nothinking"
            search_nothinking["displayName"] = model["displayName"] + " with Google Search (No Thinking)"
            search_nothinking["description"] = model["description"] + " (includes Google Search grounding, thinking disabled)"
            combined_models.append(search_nothinking)
            
            # search + maxthinking
            search_maxthinking = model.copy()
            search_maxthinking["name"] = model["name"] + "-search-maxthinking"
            search_maxthinking["displayName"] = model["displayName"] + " with Google Search (Max Thinking)"
            search_maxthinking["description"] = model["description"] + " (includes Google Search grounding, maximum thinking budget)"
            combined_models.append(search_maxthinking)
    return combined_models

# Supported Models (includes base models, search variants, and thinking variants)
# Combine all models and then sort them by name to group variants together
all_models = BASE_MODELS + _generate_search_variants() + _generate_thinking_variants()
SUPPORTED_MODELS = sorted(all_models, key=lambda x: x['name'])

# Helper function to get base model name from any variant
def get_base_model_name(model_name):
    """Convert variant model name to base model name."""
    # Remove all possible suffixes in order
    suffixes = ["-maxthinking", "-nothinking", "-search"]
    for suffix in suffixes:
        if model_name.endswith(suffix):
            return model_name[:-len(suffix)]
    return model_name

# Helper function to check if model uses search grounding
def is_search_model(model_name):
    """Check if model name indicates search grounding should be enabled."""
    return "-search" in model_name

# Helper function to check if model uses no thinking
def is_nothinking_model(model_name):
    """Check if model name indicates thinking should be disabled."""
    return "-nothinking" in model_name

# Helper function to check if model uses max thinking
def is_maxthinking_model(model_name):
    """Check if model name indicates maximum thinking budget should be used."""
    return "-maxthinking" in model_name

# Helper function to get thinking budget for a model
def get_thinking_budget(model_name):
    """Get the appropriate thinking budget for a model based on its name and variant."""
    base_model = get_base_model_name(model_name)
    
    if is_nothinking_model(model_name):
        if "gemini-2.5-flash" in base_model:
            return 0  # No thinking for flash
        elif "gemini-2.5-pro" in base_model:
            return 128  # Limited thinking for pro
    elif is_maxthinking_model(model_name):
        if "gemini-2.5-flash" in base_model:
            return 24576
        elif "gemini-2.5-pro" in base_model:
            return 32768
    else:
        # Default thinking budget for regular models
        return -1  # Default for all models

# Helper function to check if thinking should be included in output
def should_include_thoughts(model_name):
    """Check if thoughts should be included in the response."""
    if is_nothinking_model(model_name):
        # For nothinking mode, still include thoughts if it's a pro model
        base_model = get_base_model_name(model_name)
        return "gemini-2.5-pro" in base_model
    else:
        # For all other modes, include thoughts
        return True