Fix bugs, improve code clarity, and enhance overall reliability across several files. (google-gemini#339)

Faisal-Alsrheed · MarkDaoust · web-flow · commit 51d806d7c9f1 · 2024-05-17T11:52:46.000-07:00
* Fix and improve

* Fix  `_make_grounding_passages` , `_make_generate_answer_request`

* fix get_default_permission_client and get_default_permission_async_client

* Add how to test all in CONTRIBUTING.md

* fix back support for `tunedModels/` in `get_model` function

* Add pytest to CONTRIBUTING.md

* Break down test_generate_text for better debugging.

* Add pip install nose2 to CONTRIBUTING.md

* Format

Change-Id: I4e222f3e01cb8d350ae293b35a88fd5f718fe3dc

* fix sloppy types in tests

Change-Id: I3ad717ca26e5d170e4bbef23076e528badaaaacb

* Update CONTRIBUTING.md

* Update CONTRIBUTING.md

---------

Co-authored-by: Mark Daoust &lt;markdaoust@google.com&gt;
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -62,17 +62,41 @@ This "editable" mode lets you edit the source without needing to reinstall the p
 
 ### Testing
 
-Use the builtin unittest package:
+To ensure the integrity of the codebase, we have a suite of tests located in the `generative-ai-python/tests` directory. 
 
+You can run all these tests using Python's built-in `unittest` module or the `pytest` library. 
+
+For `unittest`, open a terminal and navigate to the root directory of the project. Then, execute the following command:
+
+```
+python -m unittest discover -s tests
+
+# or more simply
+python -m unittest
 ```
- python -m unittest
+
+Alternatively, if you prefer using `pytest`, you can install it using pip:
+
 ```
+pip install pytest
+```
+
+Then, run the tests with the following command:
+
+```
+pytest tests
+
+# or more simply
+pytest
+```
+
 
 Or to debug, use:
 
 ```commandline
+pip install nose2
+
 nose2 --debugger
-```
 
 ### Type checking 
 
diff --git a/google/generativeai/answer.py b/google/generativeai/answer.py
@@ -94,7 +94,7 @@ def _make_grounding_passages(source: GroundingPassagesOptions) -> glm.GroundingP
 
     if not isinstance(source, Iterable):
         raise TypeError(
-            f"`source` must be a valid `GroundingPassagesOptions` type object got a: `{type(source)}`."
+            f"The 'source' argument must be an instance of 'GroundingPassagesOptions', but got a '{type(source).__name__}' object instead."
         )
 
     passages = []
@@ -182,7 +182,7 @@ def _make_generate_answer_request(
     temperature: float | None = None,
 ) -> glm.GenerateAnswerRequest:
     """
-    Calls the API to generate a grounded answer from the model.
+    constructs a glm.GenerateAnswerRequest object by organizing the input parameters for the API call to generate a grounded answer from the model.
 
     Args:
         model: Name of the model used to generate the grounded response.
@@ -217,7 +217,7 @@ def _make_generate_answer_request(
     elif semantic_retriever is not None:
         semantic_retriever = _make_semantic_retriever_config(semantic_retriever, contents[-1])
     else:
-        TypeError(
+        raise TypeError(
             f"The source must be either an `inline_passages` xor `semantic_retriever_config`, but both are `None`"
         )
 
diff --git a/google/generativeai/client.py b/google/generativeai/client.py
@@ -328,9 +328,9 @@ def get_default_retriever_async_client() -> glm.RetrieverAsyncClient:
     return _client_manager.get_default_client("retriever_async")
 
 
-def get_dafault_permission_client() -> glm.PermissionServiceClient:
+def get_default_permission_client() -> glm.PermissionServiceClient:
     return _client_manager.get_default_client("permission")
 
 
-def get_dafault_permission_async_client() -> glm.PermissionServiceAsyncClient:
+def get_default_permission_async_client() -> glm.PermissionServiceAsyncClient:
     return _client_manager.get_default_client("permission_async")
diff --git a/google/generativeai/generative_models.py b/google/generativeai/generative_models.py
@@ -387,7 +387,7 @@ def start_chat(
         >>> response = chat.send_message("Hello?")
 
         Arguments:
-            history: An iterable of `glm.Content` objects, or equvalents to initialize the session.
+            history: An iterable of `glm.Content` objects, or equivalents to initialize the session.
         """
         if self._generation_config.get("candidate_count", 1) > 1:
             raise ValueError("Can't chat with `candidate_count > 1`")
@@ -401,11 +401,13 @@ def start_chat(
 class ChatSession:
     """Contains an ongoing conversation with the model.
 
-    >>> model = genai.GenerativeModel(model="gemini-pro")
+    >>> model = genai.GenerativeModel('models/gemini-pro')
     >>> chat = model.start_chat()
     >>> response = chat.send_message("Hello")
     >>> print(response.text)
-    >>> response = chat.send_message(...)
+    >>> response = chat.send_message("Hello again")
+    >>> print(response.text)
+    >>> response = chat.send_message(...
 
     This `ChatSession` object collects the messages sent and received, in its
     `ChatSession.history` attribute.
@@ -444,7 +446,7 @@ def send_message(
 
         Appends the request and response to the conversation history.
 
-        >>> model = genai.GenerativeModel(model="gemini-pro")
+        >>> model = genai.GenerativeModel('models/gemini-pro')
         >>> chat = model.start_chat()
         >>> response = chat.send_message("Hello")
         >>> print(response.text)
diff --git a/google/generativeai/models.py b/google/generativeai/models.py
@@ -33,29 +33,31 @@ def get_model(
     client=None,
     request_options: dict[str, Any] | None = None,
 ) -> model_types.Model | model_types.TunedModel:
-    """Given a model name, fetch the `types.Model` or `types.TunedModel` object.
+    """Given a model name, fetch the `types.Model`
 
     ```
     import pprint
-    model = genai.get_tuned_model(model_name):
+    model = genai.get_model('models/gemini-pro')
     pprint.pprint(model)
     ```
 
     Args:
-        name: The name of the model to fetch.
+        name: The name of the model to fetch. Should start with `models/`
         client: The client to use.
         request_options: Options for the request.
 
     Returns:
-        A `types.Model` or `types.TunedModel` object.
+        A `types.Model`
     """
     name = model_types.make_model_name(name)
     if name.startswith("models/"):
         return get_base_model(name, client=client, request_options=request_options)
     elif name.startswith("tunedModels/"):
         return get_tuned_model(name, client=client, request_options=request_options)
     else:
-        raise ValueError("Model names must start with `models/` or `tunedModels/`")
+        raise ValueError(
+            f"Model names must start with `models/` or `tunedModels/`. Received: {name}"
+        )
 
 
 def get_base_model(
@@ -68,12 +70,12 @@ def get_base_model(
 
     ```
     import pprint
-    model = genai.get_model('models/chat-bison-001'):
+    model = genai.get_base_model('models/chat-bison-001')
     pprint.pprint(model)
     ```
 
     Args:
-        name: The name of the model to fetch.
+        name: The name of the model to fetch. Should start with `models/`
         client: The client to use.
         request_options: Options for the request.
 
@@ -88,7 +90,7 @@ def get_base_model(
 
     name = model_types.make_model_name(name)
     if not name.startswith("models/"):
-        raise ValueError(f"Base model names must start with `models/`, got: {name}")
+        raise ValueError(f"Base model names must start with `models/`, received: {name}")
 
     result = client.get_model(name=name, **request_options)
     result = type(result).to_dict(result)
@@ -105,12 +107,12 @@ def get_tuned_model(
 
     ```
     import pprint
-    model = genai.get_tuned_model('tunedModels/my-model-1234'):
+    model = genai.get_tuned_model('tunedModels/gemini-1.0-pro-001')
     pprint.pprint(model)
     ```
 
     Args:
-        name: The name of the model to fetch.
+        name: The name of the model to fetch. Should start with `tunedModels/`
         client: The client to use.
         request_options: Options for the request.
 
@@ -126,7 +128,7 @@ def get_tuned_model(
     name = model_types.make_model_name(name)
 
     if not name.startswith("tunedModels/"):
-        raise ValueError("Tuned model names must start with `tunedModels/`")
+        raise ValueError("Tuned model names must start with `tunedModels/` received: {name}")
 
     result = client.get_tuned_model(name=name, **request_options)
 
diff --git a/google/generativeai/types/permission_types.py b/google/generativeai/types/permission_types.py
@@ -22,8 +22,8 @@
 
 from google.protobuf import field_mask_pb2
 
-from google.generativeai.client import get_dafault_permission_client
-from google.generativeai.client import get_dafault_permission_async_client
+from google.generativeai.client import get_default_permission_client
+from google.generativeai.client import get_default_permission_async_client
 from google.generativeai.utils import flatten_update_paths
 from google.generativeai import string_utils
 
@@ -107,7 +107,7 @@ def delete(
         Delete permission (self).
         """
         if client is None:
-            client = get_dafault_permission_client()
+            client = get_default_permission_client()
         delete_request = glm.DeletePermissionRequest(name=self.name)
         client.delete_permission(request=delete_request)
 
@@ -119,7 +119,7 @@ async def delete_async(
         This is the async version of `Permission.delete`.
         """
         if client is None:
-            client = get_dafault_permission_async_client()
+            client = get_default_permission_async_client()
         delete_request = glm.DeletePermissionRequest(name=self.name)
         await client.delete_permission(request=delete_request)
 
@@ -146,7 +146,7 @@ def update(
             `Permission` object with specified updates.
         """
         if client is None:
-            client = get_dafault_permission_client()
+            client = get_default_permission_client()
 
         updates = flatten_update_paths(updates)
         for update_path in updates:
@@ -176,7 +176,7 @@ async def update_async(
         This is the async version of `Permission.update`.
         """
         if client is None:
-            client = get_dafault_permission_async_client()
+            client = get_default_permission_async_client()
 
         updates = flatten_update_paths(updates)
         for update_path in updates:
@@ -224,7 +224,7 @@ def get(
             Requested permission as an instance of `Permission`.
         """
         if client is None:
-            client = get_dafault_permission_client()
+            client = get_default_permission_client()
         get_perm_request = glm.GetPermissionRequest(name=name)
         get_perm_response = client.get_permission(request=get_perm_request)
         get_perm_response = type(get_perm_response).to_dict(get_perm_response)
@@ -240,7 +240,7 @@ async def get_async(
         This is the async version of `Permission.get`.
         """
         if client is None:
-            client = get_dafault_permission_async_client()
+            client = get_default_permission_async_client()
         get_perm_request = glm.GetPermissionRequest(name=name)
         get_perm_response = await client.get_permission(request=get_perm_request)
         get_perm_response = type(get_perm_response).to_dict(get_perm_response)
@@ -313,7 +313,7 @@ def create(
             ValueError: When email_address is not specified and grantee_type is not set to EVERYONE.
         """
         if client is None:
-            client = get_dafault_permission_client()
+            client = get_default_permission_client()
 
         request = self._make_create_permission_request(
             role=role, grantee_type=grantee_type, email_address=email_address
@@ -333,7 +333,7 @@ async def create_async(
         This is the async version of `PermissionAdapter.create_permission`.
         """
         if client is None:
-            client = get_dafault_permission_async_client()
+            client = get_default_permission_async_client()
 
         request = self._make_create_permission_request(
             role=role, grantee_type=grantee_type, email_address=email_address
@@ -358,7 +358,7 @@ def list(
             Paginated list of `Permission` objects.
         """
         if client is None:
-            client = get_dafault_permission_client()
+            client = get_default_permission_client()
 
         request = glm.ListPermissionsRequest(
             parent=self.parent, page_size=page_size  # pytype: disable=attribute-error
@@ -376,7 +376,7 @@ async def list_async(
         This is the async version of `PermissionAdapter.list_permissions`.
         """
         if client is None:
-            client = get_dafault_permission_async_client()
+            client = get_default_permission_async_client()
 
         request = glm.ListPermissionsRequest(
             parent=self.parent, page_size=page_size  # pytype: disable=attribute-error
@@ -400,7 +400,7 @@ def transfer_ownership(
         if self.parent.startswith("corpora"):
             raise NotImplementedError("Can'/t transfer_ownership for a Corpus")
         if client is None:
-            client = get_dafault_permission_client()
+            client = get_default_permission_client()
         transfer_request = glm.TransferOwnershipRequest(
             name=self.parent, email_address=email_address  # pytype: disable=attribute-error
         )
@@ -415,7 +415,7 @@ async def transfer_ownership_async(
         if self.parent.startswith("corpora"):
             raise NotImplementedError("Can'/t transfer_ownership for a Corpus")
         if client is None:
-            client = get_dafault_permission_async_client()
+            client = get_default_permission_async_client()
         transfer_request = glm.TransferOwnershipRequest(
             name=self.parent, email_address=email_address  # pytype: disable=attribute-error
         )
diff --git a/tests/notebook/text_model_test.py b/tests/notebook/text_model_test.py
@@ -68,21 +68,47 @@ def _generate_text(
 
 
 class TextModelTestCase(absltest.TestCase):
-    def test_generate_text(self):
+    def test_generate_text_without_args(self):
         model = TestModel()
 
         result = model.call_model("prompt goes in")
         self.assertEqual(result.text_results[0], "prompt goes in_1")
-        self.assertIsNone(result.text_results[1])
-        self.assertIsNone(result.text_results[2])
-        self.assertIsNone(result.text_results[3])
 
+    def test_generate_text_without_args_none_results(self):
+        model = TestModel()
+
+        result = model.call_model("prompt goes in")
+        self.assertEqual(result.text_results[1], "None")
+        self.assertEqual(result.text_results[2], "None")
+        self.assertEqual(result.text_results[3], "None")
+
+    def test_generate_text_with_args_first_result(self):
+        model = TestModel()
         args = model_lib.ModelArguments(model="model_name", temperature=0.42, candidate_count=5)
+
         result = model.call_model("prompt goes in", args)
         self.assertEqual(result.text_results[0], "prompt goes in_1")
+
+    def test_generate_text_with_args_model_name(self):
+        model = TestModel()
+        args = model_lib.ModelArguments(model="model_name", temperature=0.42, candidate_count=5)
+
+        result = model.call_model("prompt goes in", args)
         self.assertEqual(result.text_results[1], "model_name")
-        self.assertEqual(result.text_results[2], 0.42)
-        self.assertEqual(result.text_results[3], 5)
+
+    def test_generate_text_with_args_temperature(self):
+        model = TestModel()
+        args = model_lib.ModelArguments(model="model_name", temperature=0.42, candidate_count=5)
+        result = model.call_model("prompt goes in", args)
+
+        self.assertEqual(result.text_results[2], str(0.42))
+
+    def test_generate_text_with_args_candidate_count(self):
+        model = TestModel()
+        args = model_lib.ModelArguments(model="model_name", temperature=0.42, candidate_count=5)
+
+        result = model.call_model("prompt goes in", args)
+        self.assertEqual(result.text_results[3], str(5))
 
     def test_retry(self):
         model = TestModel()
