llm-gemini 0.20a2__py3-none-any.whl → 0.21__py3-none-any.whl

{llm_gemini-0.20a2.dist-info → llm_gemini-0.21.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: llm-gemini
-Version: 0.20a2
+Version: 0.21
 Summary: LLM plugin to access Google's Gemini family of models
 Author: Simon Willison
 License-Expression: Apache-2.0
@@ -10,7 +10,7 @@ Project-URL: Issues, https://github.com/simonw/llm-gemini/issues
 Project-URL: CI, https://github.com/simonw/llm-gemini/actions
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: llm>=0.26a0
+Requires-Dist: llm>=0.26
 Requires-Dist: httpx
 Requires-Dist: ijson
 Provides-Extra: test
@@ -18,6 +18,7 @@ Requires-Dist: pytest; extra == "test"
 Requires-Dist: pytest-recording; extra == "test"
 Requires-Dist: pytest-asyncio; extra == "test"
 Requires-Dist: nest-asyncio; extra == "test"
+Requires-Dist: cogapp; extra == "test"
 Dynamic: license-file
 
 # llm-gemini
@@ -63,26 +64,71 @@ llm models default gemini-2.0-flash
 llm "A joke about a pelican and a walrus"
 ```
 
-Other models are:
-
-- `gemini-2.5-pro-preview-05-06` - latest paid Gemini 2.5 Pro preview
-- `gemini-2.5-flash-preview-05-20` - Gemini 2.5 Flash preview
-- `gemini-2.5-flash-preview-04-17` - Earlier Gemini 2.5 Flash preview
-- `gemini-2.5-pro-exp-03-25` - free experimental release of Gemini 2.5 Pro
-- `gemini-2.5-pro-preview-03-25` - paid preview of Gemini 2.5 Pro
-- `gemma-3-27b-it` - [Gemma 3](https://blog.google/technology/developers/gemma-3/) 27B
-- `gemini-2.0-pro-exp-02-05` - experimental release of Gemini 2.0 Pro
-- `gemini-2.0-flash-lite` - Gemini 2.0 Flash-Lite
-- `gemini-2.0-flash` - Gemini 2.0 Flash
-- `gemini-2.0-flash-thinking-exp-01-21` - experimental "thinking" model from January 2025
-- `gemini-2.0-flash-thinking-exp-1219` - experimental "thinking" model from December 2024
-- `learnlm-1.5-pro-experimental` - "an experimental task-specific model that has been trained to align with learning science principles" - [more details here](https://ai.google.dev/gemini-api/docs/learnlm).
-- `gemini-2.0-flash-exp` - [Gemini 2.0 Flash](https://blog.google/technology/google-deepmind/google-gemini-ai-update-december-2024/#gemini-2-0-flash)
-- `gemini-exp-1206` - recent experimental #3
-- `gemini-exp-1121` - recent experimental #2
-- `gemini-exp-1114` - recent experimental #1
-- `gemini-1.5-flash-8b-latest` - the least expensive
-- `gemini-1.5-flash-latest`
+## Available models
+
+<!-- [[[cog
+import cog
+from llm import cli
+from click.testing import CliRunner
+runner = CliRunner()
+result = runner.invoke(cli.cli, ["models", "-q", "gemini/"])
+lines = reversed(result.output.strip().split("\n"))
+to_output = []
+NOTES = {
+    "gemini/gemini-2.5-pro-preview-05-06": "Latest paid Gemini 2.5 Pro preview",
+    "gemini/gemini-2.5-flash-preview-05-20": "Gemini 2.5 Flash preview",
+    "gemini/gemini-2.5-flash-preview-04-17": "Earlier Gemini 2.5 Flash preview",
+    "gemini/gemini-2.5-pro-exp-03-25": "Free experimental release of Gemini 2.5 Pro",
+    "gemini/gemini-2.0-flash-thinking-exp-01-21": "Experimental \"thinking\" model from January 2025",
+    "gemini/gemini-1.5-flash-8b-latest": "The least expensive model",
+}
+for line in lines:
+    model_id, rest = line.split(None, 2)[1:]
+    note = NOTES.get(model_id, "")
+    to_output.append(
+        "- `{}`{}".format(
+            model_id,
+            ': {}'.format(note) if note else ""
+        )
+    )
+cog.out("\n".join(to_output))
+]]] -->
+- `gemini/gemini-2.5-flash-preview-05-20`: Gemini 2.5 Flash preview
+- `gemini/gemini-2.5-pro-preview-05-06`: Latest paid Gemini 2.5 Pro preview
+- `gemini/gemini-2.5-flash-preview-04-17`: Earlier Gemini 2.5 Flash preview
+- `gemini/gemini-2.5-pro-preview-03-25`
+- `gemini/gemini-2.5-pro-exp-03-25`: Free experimental release of Gemini 2.5 Pro
+- `gemini/gemini-2.0-flash-lite`
+- `gemini/gemini-2.0-pro-exp-02-05`
+- `gemini/gemini-2.0-flash`
+- `gemini/gemini-2.0-flash-thinking-exp-01-21`: Experimental "thinking" model from January 2025
+- `gemini/gemini-2.0-flash-thinking-exp-1219`
+- `gemini/gemma-3n-e4b-it`
+- `gemini/gemma-3-27b-it`
+- `gemini/gemma-3-12b-it`
+- `gemini/gemma-3-4b-it`
+- `gemini/gemma-3-1b-it`
+- `gemini/learnlm-1.5-pro-experimental`
+- `gemini/gemini-2.0-flash-exp`
+- `gemini/gemini-exp-1206`
+- `gemini/gemini-exp-1121`
+- `gemini/gemini-exp-1114`
+- `gemini/gemini-1.5-flash-8b-001`
+- `gemini/gemini-1.5-flash-8b-latest`: The least expensive model
+- `gemini/gemini-1.5-flash-002`
+- `gemini/gemini-1.5-pro-002`
+- `gemini/gemini-1.5-flash-001`
+- `gemini/gemini-1.5-pro-001`
+- `gemini/gemini-1.5-flash-latest`
+- `gemini/gemini-1.5-pro-latest`
+- `gemini/gemini-pro`
+<!-- [[[end]]] -->
+
+All of these models have aliases that omit the `gemini/` prefix, for example:
+
+```bash
+llm -m gemini-1.5-flash-8b-latest --schema 'name,age int,bio' 'invent a dog'
+```
 
 ### Images, audio and video
 
@@ -155,6 +201,31 @@ To chat interactively with the model, run `llm chat`:
 llm chat -m gemini-2.0-flash
 ```
 
+### Timeouts
+
+By default there is no `timeout` against the Gemini API. You can use the `timeout` option to protect against API requests that hang indefinitely.
+
+With the CLI tool that looks like this, to set a 1.5 second timeout:
+
+```bash
+llm -m gemini-2.5-flash-preview-05-20 'epic saga about mice' -o timeout 1.5
+```
+In the Python library timeouts are used like this:
+```python
+import httpx, llm
+
+model = llm.get_model("gemini/gemini-2.5-flash-preview-05-20")
+
+try:
+    response = model.prompt(
+        "epic saga about mice", timeout=1.5
+    )
+    print(response.text())
+except httpx.TimeoutException:
+    print("Timeout exceeded")
+```
+An `httpx.TimeoutException` subclass will be raised if the timeout is exceeded.
+
 ## Embeddings
 
 The plugin also adds support for the `gemini-embedding-exp-03-07` and `text-embedding-004` embedding models.
@@ -187,6 +258,21 @@ llm similar readmes -c 'upload csvs to stuff' -d embed.db
 
 See the [LLM embeddings documentation](https://llm.datasette.io/en/stable/embeddings/cli.html) for further details.
 
+## Listing all Gemini API models
+
+The `llm gemini models` command lists all of the models that are exposed by the Gemini API, some of which may not be available through this plugin.
+
+```bash
+llm gemini models
+```
+You can add a `--key X` option to use a different API key.
+
+To filter models by their supported generation methods use `--method` one or more times:
+```bash
+llm gemini models --method embedContent
+```
+If you provide multiple methods you will see models that support any of them.
+
 ## Development
 
 To set up this plugin locally, first checkout the code. Then create a new virtual environment:

llm_gemini-0.21.dist-info/RECORD

@@ -0,0 +1,7 @@
+llm_gemini.py,sha256=UX6G2_33rQ6_XH0v8fR3eUugTQGDOfTb6FUmAWO_fZI,23072
+llm_gemini-0.21.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+llm_gemini-0.21.dist-info/METADATA,sha256=C7C0De124aFPc_2eg0o_wr15akt1KJnuS5suiHLB9p4,10416
+llm_gemini-0.21.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+llm_gemini-0.21.dist-info/entry_points.txt,sha256=n544bpgUPIBc5l_cnwsTxPc3gMGJHPtAyqBNp-CkMWk,26
+llm_gemini-0.21.dist-info/top_level.txt,sha256=WUQmG6_2QKbT_8W4HH93qyKl_0SUteL4Ra6_PhyNGKU,11
+llm_gemini-0.21.dist-info/RECORD,,

{llm_gemini-0.20a2.dist-info → llm_gemini-0.21.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: setuptools (80.8.0)
+Generator: setuptools (80.9.0)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

llm_gemini.py

@@ -55,15 +55,56 @@ GOOGLE_SEARCH_MODELS_USING_SEARCH_RETRIEVAL = {
 }
 
 THINKING_BUDGET_MODELS = {
+    "gemini-2.0-flash-thinking-exp-01-21",
+    "gemini-2.0-flash-thinking-exp-1219",
     "gemini-2.5-flash-preview-04-17",
+    "gemini-2.5-pro-exp-03-25",
+    "gemini-2.5-pro-preview-03-25",
+    "gemini-2.5-pro-preview-05-06",
     "gemini-2.5-flash-preview-05-20",
 }
 
+NO_VISION_MODELS = {"gemma-3-1b-it", "gemma-3n-e4b-it"}
+
+ATTACHMENT_TYPES = {
+    # Text
+    "text/plain",
+    "text/csv",
+    # PDF
+    "application/pdf",
+    # Images
+    "image/png",
+    "image/jpeg",
+    "image/webp",
+    "image/heic",
+    "image/heif",
+    # Audio
+    "audio/wav",
+    "audio/mp3",
+    "audio/aiff",
+    "audio/aac",
+    "audio/ogg",
+    "application/ogg",
+    "audio/flac",
+    "audio/mpeg",  # Treated as audio/mp3
+    # Video
+    "video/mp4",
+    "video/mpeg",
+    "video/mov",
+    "video/avi",
+    "video/x-flv",
+    "video/mpg",
+    "video/webm",
+    "video/wmv",
+    "video/3gpp",
+    "video/quicktime",
+}
+
 
 @llm.hookimpl
 def register_models(register):
     # Register both sync and async versions of each model
-    for model_id in [
+    for model_id in (
         "gemini-pro",
         "gemini-1.5-pro-latest",
         "gemini-1.5-flash-latest",
@@ -78,6 +119,12 @@ def register_models(register):
         "gemini-exp-1206",
         "gemini-2.0-flash-exp",
         "learnlm-1.5-pro-experimental",
+        # Gemma 3 models:
+        "gemma-3-1b-it",
+        "gemma-3-4b-it",
+        "gemma-3-12b-it",  # 12th March 2025
+        "gemma-3-27b-it",
+        "gemma-3n-e4b-it",  # 20th May 2025
         "gemini-2.0-flash-thinking-exp-1219",
         "gemini-2.0-flash-thinking-exp-01-21",
         # Released 5th Feb 2025:
@@ -85,8 +132,6 @@ def register_models(register):
         "gemini-2.0-pro-exp-02-05",
         # Released 25th Feb 2025:
         "gemini-2.0-flash-lite",
-        # Released 12th March 2025:
-        "gemma-3-27b-it",
         # 25th March 2025:
         "gemini-2.5-pro-exp-03-25",
         # 4th April 2025 (paid):
@@ -97,22 +142,27 @@ def register_models(register):
         "gemini-2.5-pro-preview-05-06",
         # 20th May 2025:
         "gemini-2.5-flash-preview-05-20",
-    ]:
+    ):
         can_google_search = model_id in GOOGLE_SEARCH_MODELS
         can_thinking_budget = model_id in THINKING_BUDGET_MODELS
+        can_vision = model_id not in NO_VISION_MODELS
+        can_schema = "flash-thinking" not in model_id and "gemma-3" not in model_id
         register(
             GeminiPro(
                 model_id,
+                can_vision=can_vision,
                 can_google_search=can_google_search,
                 can_thinking_budget=can_thinking_budget,
-                can_schema="flash-thinking" not in model_id,
+                can_schema=can_schema,
             ),
             AsyncGeminiPro(
                 model_id,
+                can_vision=can_vision,
                 can_google_search=can_google_search,
                 can_thinking_budget=can_thinking_budget,
-                can_schema="flash-thinking" not in model_id,
+                can_schema=can_schema,
             ),
+            aliases=(model_id,),
         )
 
 
@@ -154,39 +204,7 @@ class _SharedGemini:
     supports_schema = True
     supports_tools = True
 
-    attachment_types = (
-        # Text
-        "text/plain",
-        "text/csv",
-        # PDF
-        "application/pdf",
-        # Images
-        "image/png",
-        "image/jpeg",
-        "image/webp",
-        "image/heic",
-        "image/heif",
-        # Audio
-        "audio/wav",
-        "audio/mp3",
-        "audio/aiff",
-        "audio/aac",
-        "audio/ogg",
-        "application/ogg",
-        "audio/flac",
-        "audio/mpeg",  # Treated as audio/mp3
-        # Video
-        "video/mp4",
-        "video/mpeg",
-        "video/mov",
-        "video/avi",
-        "video/x-flv",
-        "video/mpg",
-        "video/webm",
-        "video/wmv",
-        "video/3gpp",
-        "video/quicktime",
-    )
+    attachment_types = set()
 
     class Options(llm.Options):
         code_execution: Optional[bool] = Field(
@@ -232,6 +250,14 @@ class _SharedGemini:
             description="Output a valid JSON object {...}",
             default=None,
         )
+        timeout: Optional[float] = Field(
+            description=(
+                "The maximum time in seconds to wait for a response. "
+                "If the model does not respond within this time, "
+                "the request will be aborted."
+            ),
+            default=None,
+        )
 
     class OptionsWithGoogleSearch(Options):
         google_search: Optional[bool] = Field(
@@ -247,12 +273,14 @@ class _SharedGemini:
 
     def __init__(
         self,
-        model_id,
+        gemini_model_id,
+        can_vision=True,
         can_google_search=False,
         can_thinking_budget=False,
         can_schema=False,
     ):
-        self.model_id = model_id
+        self.model_id = "gemini/{}".format(gemini_model_id)
+        self.gemini_model_id = gemini_model_id
         self.can_google_search = can_google_search
         self.supports_schema = can_schema
         if can_google_search:
@@ -260,6 +288,8 @@ class _SharedGemini:
         self.can_thinking_budget = can_thinking_budget
         if can_thinking_budget:
             self.Options = self.OptionsWithThinkingBudget
+        if can_vision:
+            self.attachment_types = ATTACHMENT_TYPES
 
     def build_messages(self, prompt, conversation):
         messages = []
@@ -457,14 +487,14 @@ class _SharedGemini:
 
 class GeminiPro(_SharedGemini, llm.KeyModel):
     def execute(self, prompt, stream, response, conversation, key):
-        url = f"https://generativelanguage.googleapis.com/v1beta/models/{self.model_id}:streamGenerateContent"
+        url = f"https://generativelanguage.googleapis.com/v1beta/models/{self.gemini_model_id}:streamGenerateContent"
         gathered = []
         body = self.build_request_body(prompt, conversation)
 
         with httpx.stream(
             "POST",
             url,
-            timeout=None,
+            timeout=prompt.options.timeout,
             headers={"x-goog-api-key": self.get_key(key)},
             json=body,
         ) as http_response:
@@ -490,7 +520,7 @@ class GeminiPro(_SharedGemini, llm.KeyModel):
 
 class AsyncGeminiPro(_SharedGemini, llm.AsyncKeyModel):
     async def execute(self, prompt, stream, response, conversation, key):
-        url = f"https://generativelanguage.googleapis.com/v1beta/models/{self.model_id}:streamGenerateContent"
+        url = f"https://generativelanguage.googleapis.com/v1beta/models/{self.gemini_model_id}:streamGenerateContent"
         gathered = []
         body = self.build_request_body(prompt, conversation)
 
@@ -498,7 +528,7 @@ class AsyncGeminiPro(_SharedGemini, llm.AsyncKeyModel):
             async with client.stream(
                 "POST",
                 url,
-                timeout=None,
+                timeout=prompt.options.timeout,
                 headers={"x-goog-api-key": self.get_key(key)},
                 json=body,
             ) as http_response:
@@ -588,8 +618,20 @@ def register_commands(cli):
 
     @gemini.command()
     @click.option("--key", help="API key to use")
-    def models(key):
-        "List of Gemini models pulled from their API"
+    @click.option(
+        "methods",
+        "--method",
+        multiple=True,
+        help="Filter by supported generation methods",
+    )
+    def models(key, methods):
+        """
+        List of Gemini models pulled from their API
+
+        Use --method to filter by supported generation methods for example:
+
+        llm gemini models --method generateContent --method embedContent
+        """
         key = llm.get_key(key, "gemini", "LLM_GEMINI_KEY")
         if not key:
             raise click.ClickException(
@@ -598,7 +640,16 @@ def register_commands(cli):
         url = f"https://generativelanguage.googleapis.com/v1beta/models"
         response = httpx.get(url, headers={"x-goog-api-key": key})
         response.raise_for_status()
-        click.echo(json.dumps(response.json()["models"], indent=2))
+        models = response.json()["models"]
+        if methods:
+            models = [
+                model
+                for model in models
+                if any(
+                    method in model["supportedGenerationMethods"] for method in methods
+                )
+            ]
+        click.echo(json.dumps(models, indent=2))
 
     @gemini.command()
     @click.option("--key", help="API key to use")

llm_gemini-0.20a2.dist-info/RECORD

@@ -1,7 +0,0 @@
-llm_gemini.py,sha256=rHA904SF-06hjTf2bem219ZxK98DlS6BuUVLl92FffM,21465
-llm_gemini-0.20a2.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-llm_gemini-0.20a2.dist-info/METADATA,sha256=gQD_OgG2F6UP3Lpvf-TsfVIlCCx0djUualXyIwC24Bk,8139
-llm_gemini-0.20a2.dist-info/WHEEL,sha256=zaaOINJESkSfm_4HQVc5ssNzHCPXhJm0kEUakpsEHaU,91
-llm_gemini-0.20a2.dist-info/entry_points.txt,sha256=n544bpgUPIBc5l_cnwsTxPc3gMGJHPtAyqBNp-CkMWk,26
-llm_gemini-0.20a2.dist-info/top_level.txt,sha256=WUQmG6_2QKbT_8W4HH93qyKl_0SUteL4Ra6_PhyNGKU,11
-llm_gemini-0.20a2.dist-info/RECORD,,