Merge pull request #994 from kristiankyvik/generate-docstrings

Looks great! If you intend to be at the meeting on Thursday, I would be curious to hear how you did this in more detail!

docs/conf.py

@@ -90,8 +90,7 @@
     "members": True,
     "show-inheritance": True,
     "inherited-members": "BaseModel",
-    "undoc-members": True,
-    "special-members": "__call__",
+    "undoc-members": False,
 }
 
 # Add any paths that contain templates here, relative to this directory.

docs/installation.rst

@@ -46,7 +46,7 @@ Once you have a copy of the source, you can install it with:
 .. _Github repo: https://github.com/gpt-engineer-org/gpt-engineer.git
 
 Troubleshooting
--------------
+---------------
 
 For mac and linux system, there are sometimes slim python installations that do not include the gpt-engineer requirement tkinter, which is a standard library and thus not pip installable.
 

gpt_engineer/applications/cli/__init__.py

@@ -1,15 +0,0 @@
-"""
-gpt_engineer.cli
------------------
-
-The CLI package for the GPT Engineer project, providing the command line interface
-for the application.
-
-Modules:
-    - main: The primary CLI module for GPT Engineer.
-    - collect: Collect send learning tools for analysis and improvement.
-    - file_selector: Selecting files using GUI and terminal-based file explorer.
-    - learning: Tools and data structures for data collection.
-
-For more specific details, refer to the docstrings within each module.
-"""

gpt_engineer/applications/cli/cli_agent.py

@@ -1,3 +1,9 @@
+"""
+This module provides the CliAgent class which manages the lifecycle of code generation and improvement
+using an AI model. It includes functionalities to initialize code generation, improve existing code,
+and process the code through various steps defined in the step bundle.
+"""
+
 from typing import Callable, Optional, TypeVar
 
 # from gpt_engineer.core.default.git_version_manager import GitVersionManager
@@ -28,43 +34,50 @@
 
 class CliAgent(BaseAgent):
     """
-    The `Agent` class is responsible for managing the lifecycle of code generation and improvement.
-
-    Attributes:
-        path (str): The file path where the `Agent` will operate, used for version management and
-                    file operations.
-        version_manager (BaseVersionManager): An object that adheres to the VersionManagerInterface,
-                        responsible for version control of the generated code. Defaults to `VersionManager`
-                        if not provided. PROBABLY GIT SHOULD BE USED IN THE DEFAULT
-        step_bundle (StepBundleInterface): Workflows of code generation steps that define the behavior of gen_code and
-        improve.
-        ai (AI): Manages calls to the LLM.
-
-    Methods:
-        __init__(self, path: str, version_manager: VersionManagerInterface = None,
-                 step_bundle: StepBundleInterface = None, ai: AI = None):
-            Initializes a new instance of the Agent class with the provided path, version manager,
-            step bundle, and AI. It falls back to default instances if specific components are not provided.
-
-        init(self, prompt: str) -> Code:
-            Generates a new piece of code using the AI and step bundle based on the provided prompt.
-            It also snapshots the generated code using the version manager.
-
-            Parameters:
-                prompt (str): A string prompt that guides the code generation process.
-
-            Returns:
-                FilesDict: An instance of the `Code` class containing the generated code.
-
-        improve(self, prompt: str) -> Code:
-            Improves an existing piece of code using the AI and step bundle based on the provided prompt.
-            It also snapshots the improved code using the version manager.
-
-            Parameters:
-                prompt (str): A string prompt that guides the code improvement process.
-
-            Returns:
-                FilesDict: An instance of the `Code` class containing the improved code.
+    The `CliAgent` class is responsible for managing the lifecycle of code generation and improvement
+    using an AI model. It orchestrates the generation of new code and the improvement of existing code
+    based on given prompts and utilizes a memory system and execution environment for processing.
+
+    Parameters
+    ----------
+    memory : BaseMemory
+        An instance of a class that adheres to the BaseMemory interface, used for storing and retrieving
+        information during the code generation process.
+    execution_env : BaseExecutionEnv
+        An instance of a class that adheres to the BaseExecutionEnv interface, used for executing code
+        and managing the execution environment.
+    ai : AI, optional
+        An instance of the AI class that manages calls to the language model. If not provided, a default
+        instance is created.
+    code_gen_fn : CodeGenType, optional
+        A callable that takes an AI instance, a prompt, and a memory instance to generate code. Defaults
+        to the `gen_code` function.
+    improve_fn : ImproveType, optional
+        A callable that takes an AI instance, a prompt, a FilesDict instance, and a memory instance to
+        improve code. Defaults to the `improve` function.
+    process_code_fn : CodeProcessor, optional
+        A callable that takes an AI instance, an execution environment, and a FilesDict instance to
+        process code. Defaults to the `execute_entrypoint` function.
+    preprompts_holder : PrepromptsHolder, optional
+        An instance of PrepromptsHolder that manages preprompt templates. If not provided, a default
+        instance is created using the PREPROMPTS_PATH.
+
+    Attributes
+    ----------
+    memory : BaseMemory
+        The memory instance where the agent stores and retrieves information.
+    execution_env : BaseExecutionEnv
+        The execution environment instance where the agent executes and manages code.
+    ai : AI
+        The AI instance used for interacting with the language model.
+    code_gen_fn : CodeGenType
+        The function used for generating code.
+    improve_fn : ImproveType
+        The function used for improving code.
+    process_code_fn : CodeProcessor
+        The function used for processing code.
+    preprompts_holder : PrepromptsHolder
+        The holder for preprompt templates.
     """
 
     def __init__(
@@ -96,6 +109,34 @@ def with_default_config(
         process_code_fn: CodeProcessor = execute_entrypoint,
         preprompts_holder: PrepromptsHolder = None,
     ):
+        """
+        Creates a new instance of CliAgent with default configurations for memory, execution environment,
+        AI, and other functional parameters.
+
+        Parameters
+        ----------
+        memory : DiskMemory
+            An instance of DiskMemory for storing and retrieving information.
+        execution_env : DiskExecutionEnv
+            An instance of DiskExecutionEnv for executing code.
+        ai : AI, optional
+            An instance of AI for interacting with the language model. Defaults to None, which will create
+            a new AI instance.
+        code_gen_fn : CodeGenType, optional
+            A function for generating code. Defaults to `gen_code`.
+        improve_fn : ImproveType, optional
+            A function for improving code. Defaults to `improve`.
+        process_code_fn : CodeProcessor, optional
+            A function for processing code. Defaults to `execute_entrypoint`.
+        preprompts_holder : PrepromptsHolder, optional
+            An instance of PrepromptsHolder for managing preprompt templates. Defaults to None, which will
+            create a new PrepromptsHolder instance using PREPROMPTS_PATH.
+
+        Returns
+        -------
+        CliAgent
+            An instance of CliAgent configured with the provided or default parameters.
+        """
         return cls(
             memory=memory,
             execution_env=execution_env,
@@ -107,6 +148,20 @@ def with_default_config(
         )
 
     def init(self, prompt: str) -> FilesDict:
+        """
+        Generates a new piece of code using the AI and step bundle based on the provided prompt.
+
+        Parameters
+        ----------
+        prompt : str
+            A string prompt that guides the code generation process.
+
+        Returns
+        -------
+        FilesDict
+            An instance of the `FilesDict` class containing the generated code.
+        """
+
         files_dict = self.code_gen_fn(
             self.ai, prompt, self.memory, self.preprompts_holder
         )
@@ -129,6 +184,24 @@ def improve(
         prompt: str,
         execution_command: Optional[str] = None,
     ) -> FilesDict:
+        """
+        Improves an existing piece of code using the AI and step bundle based on the provided prompt.
+
+        Parameters
+        ----------
+        files_dict : FilesDict
+            An instance of `FilesDict` containing the code to be improved.
+        prompt : str
+            A string prompt that guides the code improvement process.
+        execution_command : str, optional
+            An optional command to execute the code. If not provided, the default execution command is used.
+
+        Returns
+        -------
+        FilesDict
+            An instance of the `FilesDict` class containing the improved code.
+        """
+
         files_dict = self.improve_fn(
             self.ai, prompt, files_dict, self.memory, self.preprompts_holder
         )

gpt_engineer/applications/cli/collect.py

@@ -1,27 +1,25 @@
 """
+Module `collect` - Data Handling and RudderStack Integration
+
 This module provides functionalities to handle and send learning data to RudderStack
 for the purpose of analysis and to improve the gpt-engineer system. The data is sent
 only when the user gives consent to share.
 
-The module provides the following main functions:
-
-- `send_learning`: Directly send a learning data to RudderStack.
-- `collect_learnings`: Extract, possibly adjust, and send the learning data based on
-  provided input parameters.
-- `steps_file_hash`: Computes the SHA-256 hash of the steps file, which might be used
-  for identifying the exact version or changes in the steps.
+Functions:
+    send_learning(learning): Sends learning data to RudderStack.
+    collect_learnings(prompt, model, temperature, config, memory, review): Processes and sends learning data.
+    collect_and_send_human_review(prompt, model, temperature, config, memory): Collects human feedback and sends it.
 
 Dependencies:
-- hashlib: For generating SHA-256 hash.
-- typing: For type annotations.
-- gpt_engineer.core: Core functionalities of gpt-engineer.
-- gpt_engineer.cli.learning: Handles the extraction of learning data.
+    hashlib: For generating SHA-256 hash.
+    typing: For type annotations.
+    gpt_engineer.core: Core functionalities of gpt-engineer.
+    gpt_engineer.cli.learning: Handles the extraction of learning data.
 
-Note:
+Notes:
     Data sent to RudderStack is not shared with third parties and is used solely to
     improve gpt-engineer and allow it to handle a broader range of use cases.
     Consent logic is in gpt_engineer/learning.py.
-
 """
 
 from typing import Tuple
@@ -39,16 +37,17 @@ def send_learning(learning: Learning):
     """
     Send the learning data to RudderStack for analysis.
 
-    Note:
-    This function is only called if consent is given to share data.
-    Data is not shared to a third party. It is used with the sole purpose of
-    improving gpt-engineer, and letting it handle more use cases.
-    Consent logic is in gpt_engineer/learning.py
-
     Parameters
     ----------
     learning : Learning
-        The learning data to send.
+        An instance of the Learning class containing the data to be sent.
+
+    Notes
+    -----
+    This function is only called if consent is given to share data.
+    Data is not shared to a third party. It is used with the sole purpose of
+    improving gpt-engineer, and letting it handle more use cases.
+    Consent logic is in gpt_engineer/learning.py.
     """
     import rudderstack.analytics as rudder_analytics
 
@@ -75,14 +74,23 @@ def collect_learnings(
 
     Parameters
     ----------
+    prompt : str
+        The initial prompt or question that was provided to the model.
     model : str
-        The name of the model used.
+        The name of the model used for generating the response.
     temperature : float
-        The temperature used.
-    steps : List[Step]
-        The list of steps.
-    dbs : DBs
-        The database containing the workspace.
+        The temperature setting used in the model's response generation.
+    config : any
+        Configuration parameters used for the learning session.
+    memory : DiskMemory
+        An instance of DiskMemory for storing and retrieving data.
+    review : Review
+        An instance of Review containing human feedback on the model's response.
+
+    Notes
+    -----
+    This function attempts to send the learning data to RudderStack. If the data size exceeds
+    the maximum allowed size, it trims the data and retries sending it.
     """
     learnings = extract_learning(prompt, model, temperature, config, memory, review)
     try:
@@ -137,29 +145,32 @@ def collect_and_send_human_review(
     memory: DiskMemory,
 ):
     """
-    Collects human feedback on the code and stores it in memory.
-
-    This function prompts the user for a review of the generated or improved code using the `human_review_input`
-    function. If a valid review is provided, it's serialized to JSON format and stored within the database's
-    memory under the "review" key.
-
-    Parameters:
-    - ai (AI): An instance of the AI model. Although not directly used within the function, it is kept as
-      a parameter for consistency with other functions.
-    - dbs (DBs): An instance containing the database configurations, user prompts, project metadata,
-      and memory storage. This function specifically interacts with the memory storage to save the human review.
-
-    Returns:
-    - list: Returns an empty list, indicating that there's no subsequent interaction with the LLM
-      or no further messages to be processed.
-
-    Notes:
-    - It's assumed that the `human_review_input` function handles all the interactions with the user to
-      gather feedback and returns either the feedback or None if no feedback was provided.
-    - Ensure that the database's memory has enough space or is set up correctly to store the serialized review data.
+    Collects human feedback on the code and sends it for analysis.
+
+    Parameters
+    ----------
+    prompt : str
+        The initial prompt or question that was provided to the model.
+    model : str
+        The name of the model used for generating the response.
+    temperature : float
+        The temperature setting used in the model's response generation.
+    config : Tuple[str, ...]
+        Configuration parameters used for the learning session.
+    memory : DiskMemory
+        An instance of DiskMemory for storing and retrieving data.
+
+    Returns
+    -------
+    None
+
+    Notes
+    -----
+    This function prompts the user for a review of the generated or improved code using the
+    `human_review_input` function. If a valid review is provided, it's serialized to JSON format
+    and stored within the database's memory under the "review" key.
     """
 
-    """Collects and stores human review of the code"""
     review = human_review_input()
     if review:
         collect_learnings(prompt, model, temperature, config, memory, review)