Release v0.4.4 (#4906)

.env.template

@@ -58,15 +58,19 @@ OPENAI_API_KEY=your-openai-api-key
 ## USE_AZURE - Use Azure OpenAI or not (Default: False)
 # USE_AZURE=False
 
+## AZURE_CONFIG_FILE - The path to the azure.yaml file (Default: azure.yaml)
+# AZURE_CONFIG_FILE=azure.yaml
+
+
 ################################################################################
 ### LLM MODELS
 ################################################################################
 
-## SMART_LLM_MODEL - Smart language model (Default: gpt-3.5-turbo)
-# SMART_LLM_MODEL=gpt-3.5-turbo
+## SMART_LLM - Smart language model (Default: gpt-4)
+# SMART_LLM=gpt-4
 
-## FAST_LLM_MODEL - Fast language model (Default: gpt-3.5-turbo)
-# FAST_LLM_MODEL=gpt-3.5-turbo
+## FAST_LLM - Fast language model (Default: gpt-3.5-turbo)
+# FAST_LLM=gpt-3.5-turbo
 
 ## EMBEDDING_MODEL - Model to use for creating embeddings
 # EMBEDDING_MODEL=text-embedding-ada-002

.github/CODEOWNERS

@@ -1 +1,2 @@
-.github/workflows/ @Significant-Gravitas/Auto-GPT-Source
+.github/workflows/ @Significant-Gravitas/maintainers
+autogpt/core @collijk

.github/ISSUE_TEMPLATE/1.bug.yml

@@ -140,8 +140,8 @@ body:
         ⚠️The following is OPTIONAL, please keep in mind that the log files may contain personal information such as credentials.⚠️
         
         "The log files are located in the folder 'logs' inside the main auto-gpt folder."
-        
-  - type: input
+
+  - type: textarea
     attributes:
       label: Upload Activity Log Content
       description: |
@@ -152,7 +152,7 @@ body:
     validations:
       required: false
 
-  - type: input
+  - type: textarea
     attributes:
       label: Upload Error Log Content
       description: |

.github/workflows/ci.yml

@@ -108,22 +108,27 @@ jobs:
         if: ${{ startsWith(github.event_name, 'pull_request') }}
         run: |
           cassette_branch="${{ github.event.pull_request.user.login }}-${{ github.event.pull_request.head.ref }}"
+          cassette_base_branch="${{ github.event.pull_request.base.ref }}"
           cd tests/Auto-GPT-test-cassettes
 
+          if ! git ls-remote --exit-code --heads origin $cassette_base_branch ; then
+            cassette_base_branch="master"
+          fi
+
           if git ls-remote --exit-code --heads origin $cassette_branch ; then
             git fetch origin $cassette_branch
-            git fetch origin ${{ github.event.pull_request.base.ref }}
+            git fetch origin $cassette_base_branch
 
             git checkout $cassette_branch
 
             # Pick non-conflicting cassette updates from the base branch
-            git merge --no-commit --strategy-option=ours origin/${{ github.event.pull_request.base.ref }}
+            git merge --no-commit --strategy-option=ours origin/$cassette_base_branch
             echo "Using cassettes from mirror branch '$cassette_branch'," \
-              "synced to upstream branch '${{ github.event.pull_request.base.ref }}'."
+              "synced to upstream branch '$cassette_base_branch'."
           else
             git checkout -b $cassette_branch
             echo "Branch '$cassette_branch' does not exist in cassette submodule." \
-              "Using cassettes from '${{ github.event.pull_request.base.ref }}'."
+              "Using cassettes from '$cassette_base_branch'."
           fi
 
       - name: Set up Python ${{ matrix.python-version }}

.pre-commit-config.yaml

@@ -36,7 +36,7 @@ repos:
         types: [ python ]
       - id: pytest-check
         name: pytest-check
-        entry: pytest --cov=autogpt --without-integration --without-slow-integration
+        entry: pytest --cov=autogpt tests/unit
         language: system
         pass_filenames: false
         always_run: true

BULLETIN.md

@@ -1,22 +1,29 @@
-# Website and Documentation Site 📰📖
-Check out *https://agpt.co*, the official news & updates site for Auto-GPT!
-The documentation also has a place here, at *https://docs.agpt.co*
+# QUICK LINKS 🔗
+# --------------
+🌎 *Official Website*: https://agpt.co.
+📖 *User Guide*: https://docs.agpt.co.
+👩 *Contributors Wiki*: https://github.com/Significant-Gravitas/Auto-GPT/wiki/Contributing.
 
-# For contributors 👷🏼
-Since releasing v0.3.0, whave been working on re-architecting the Auto-GPT core to make it more extensible and make room for structural performance-oriented R&D.
+# v0.4.4 RELEASE HIGHLIGHTS! 🚀
+# -----------------------------
+## GPT-4 is back!
+Following OpenAI's recent GPT-4 GA announcement, the SMART_LLM .env setting 
+now defaults to GPT-4, and Auto-GPT will use GPT-4 by default in its main loop.
 
-Check out the contribution guide on our wiki:
-https://github.com/Significant-Gravitas/Auto-GPT/wiki/Contributing
+### !! High Costs Warning !! 💰💀🚨
+GPT-4 costs ~20x more than GPT-3.5-turbo. 
+Please take note of this before using SMART_LLM. You can use `--gpt3only` 
+or `--gpt4only` to force the use of GPT-3.5-turbo or GPT-4, respectively, 
+at runtime.
 
-# 🚀 v0.4.3 Release 🚀
-We're happy to announce the 0.4.3 maintenance release, which primarily focuses on refining the LLM command execution,
-extending support for OpenAI's latest models (including the powerful GPT-3 16k model), and laying the groundwork 
-for future compatibility with OpenAI's function calling feature.
+## Re-arch v1 preview release!
+We've released a preview version of the re-arch code, under `autogpt/core`. 
+This is a major milestone for us, and we're excited to continue working on it. 
+We look forward to your feedback. Follow the process here: 
+https://github.com/Significant-Gravitas/Auto-GPT/issues/4770.
 
-Key Highlights:
-- OpenAI API Key Prompt: Auto-GPT will now courteously prompt users for their OpenAI API key, if it's not already provided.
-- Summarization Enhancements: We've optimized Auto-GPT's use of the LLM context window even further.
-- JSON Memory Reading: Support for reading memories from JSON files has been improved, resulting in enhanced task execution.
-- Deprecated commands, removed for a leaner, more performant LLM: analyze_code, write_tests, improve_code, audio_text, web_playwright, web_requests
-## Take a look at the Release Notes on Github for the full changelog!
-https://github.com/Significant-Gravitas/Auto-GPT/releases
+## Other highlights
+Other fixes include plugins regressions, Azure config and security patches.
+
+Take a look at the Release Notes on Github for the full changelog! 
+https://github.com/Significant-Gravitas/Auto-GPT/releases.

autogpt/agent/agent.py

@@ -2,6 +2,7 @@
 import signal
 import sys
 from datetime import datetime
+from pathlib import Path
 
 from colorama import Fore, Style
 
@@ -64,7 +65,7 @@ def __init__(
         ai_config: AIConfig,
         system_prompt: str,
         triggering_prompt: str,
-        workspace_directory: str,
+        workspace_directory: str | Path,
         config: Config,
     ):
         self.ai_name = ai_name
@@ -80,13 +81,11 @@ def __init__(
         self.created_at = datetime.now().strftime("%Y%m%d_%H%M%S")
         self.cycle_count = 0
         self.log_cycle_handler = LogCycleHandler()
-        self.fast_token_limit = OPEN_AI_CHAT_MODELS.get(
-            config.fast_llm_model
-        ).max_tokens
+        self.smart_token_limit = OPEN_AI_CHAT_MODELS.get(config.smart_llm).max_tokens
 
     def start_interaction_loop(self):
         # Avoid circular imports
-        from autogpt.app import execute_command, get_command
+        from autogpt.app import execute_command, extract_command
 
         # Interaction Loop
         self.cycle_count = 0
@@ -137,8 +136,8 @@ def signal_handler(signum, frame):
                     self,
                     self.system_prompt,
                     self.triggering_prompt,
-                    self.fast_token_limit,
-                    self.config.fast_llm_model,
+                    self.smart_token_limit,
+                    self.config.smart_llm,
                 )
 
             try:
@@ -162,11 +161,11 @@ def signal_handler(signum, frame):
                     print_assistant_thoughts(
                         self.ai_name, assistant_reply_json, self.config
                     )
-                    command_name, arguments = get_command(
+                    command_name, arguments = extract_command(
                         assistant_reply_json, assistant_reply, self.config
                     )
                     if self.config.speak_mode:
-                        say_text(f"I want to execute {command_name}")
+                        say_text(f"I want to execute {command_name}", self.config)
 
                     arguments = self._resolve_pathlike_command_args(arguments)
 
@@ -195,8 +194,9 @@ def signal_handler(signum, frame):
                 # to exit
                 self.user_input = ""
                 logger.info(
-                    "Enter 'y' to authorise command, 'y -N' to run N continuous commands, 's' to run self-feedback commands, "
-                    "'n' to exit program, or enter feedback for "
+                    f"Enter '{self.config.authorise_key}' to authorise command, "
+                    f"'{self.config.authorise_key} -N' to run N continuous commands, "
+                    f"'{self.config.exit_key}' to exit program, or enter feedback for "
                     f"{self.ai_name}..."
                 )
                 while True:
@@ -224,8 +224,8 @@ def signal_handler(signum, frame):
                             user_input = "GENERATE NEXT COMMAND JSON"
                         except ValueError:
                             logger.warn(
-                                "Invalid input format. Please enter 'y -n' where n is"
-                                " the number of continuous tasks."
+                                f"Invalid input format. Please enter '{self.config.authorise_key} -n' "
+                                "where n is the number of continuous tasks."
                             )
                             continue
                         break
@@ -281,12 +281,12 @@ def signal_handler(signum, frame):
                 result = f"Command {command_name} returned: " f"{command_result}"
 
                 result_tlength = count_string_tokens(
-                    str(command_result), self.config.fast_llm_model
+                    str(command_result), self.config.smart_llm
                 )
                 memory_tlength = count_string_tokens(
-                    str(self.history.summary_message()), self.config.fast_llm_model
+                    str(self.history.summary_message()), self.config.smart_llm
                 )
-                if result_tlength + memory_tlength + 600 > self.fast_token_limit:
+                if result_tlength + memory_tlength + 600 > self.smart_token_limit:
                     result = f"Failure: command {command_name} returned too much output. \
                         Do not execute this command again with the same arguments."
 

autogpt/app.py

@@ -23,7 +23,7 @@ def is_valid_int(value: str) -> bool:
         return False
 
 
-def get_command(
+def extract_command(
     assistant_reply_json: Dict, assistant_reply: ChatModelResponse, config: Config
 ):
     """Parse the response and return the command name and arguments
@@ -78,21 +78,6 @@ def get_command(
         return "Error:", str(e)
 
 
-def map_command_synonyms(command_name: str):
-    """Takes the original command name given by the AI, and checks if the
-    string matches a list of common/known hallucinations
-    """
-    synonyms = [
-        ("write_file", "write_to_file"),
-        ("create_file", "write_to_file"),
-        ("search", "google"),
-    ]
-    for seen_command, actual_command_name in synonyms:
-        if command_name == seen_command:
-            return actual_command_name
-    return command_name
-
-
 def execute_command(
     command_name: str,
     arguments: dict[str, str],
@@ -109,28 +94,21 @@ def execute_command(
         str: The result of the command
     """
     try:
-        cmd = agent.command_registry.commands.get(command_name)
+        # Execute a native command with the same name or alias, if it exists
+        if command := agent.command_registry.get_command(command_name):
+            return command(**arguments, agent=agent)
 
-        # If the command is found, call it with the provided arguments
-        if cmd:
-            return cmd(**arguments, agent=agent)
-
-        # TODO: Remove commands below after they are moved to the command registry.
-        command_name = map_command_synonyms(command_name.lower())
-
-        # TODO: Change these to take in a file rather than pasted code, if
-        # non-file is given, return instructions "Input should be a python
-        # filepath, write your code to file and try again
+        # Handle non-native commands (e.g. from plugins)
         for command in agent.ai_config.prompt_generator.commands:
             if (
                 command_name == command["label"].lower()
                 or command_name == command["name"].lower()
             ):
                 return command["function"](**arguments)
-        return (
-            f"Unknown command '{command_name}'. Please refer to the 'COMMANDS'"
-            " list for available commands and only respond in the specified JSON"
-            " format."
+
+        raise RuntimeError(
+            f"Cannot execute '{command_name}': unknown command."
+            " Do not try to use this command again."
         )
     except Exception as e:
         return f"Error: {str(e)}"

autogpt/cli.py

@@ -1,4 +1,6 @@
 """Main script for the autogpt package."""
+from typing import Optional
+
 import click
 
 
@@ -65,6 +67,22 @@
     is_flag=True,
     help="Installs external dependencies for 3rd party plugins.",
 )
+@click.option(
+    "--ai-name",
+    type=str,
+    help="AI name override",
+)
+@click.option(
+    "--ai-role",
+    type=str,
+    help="AI role override",
+)
+@click.option(
+    "--ai-goal",
+    type=str,
+    multiple=True,
+    help="AI goal override; may be used multiple times to pass multiple goals",
+)
 @click.pass_context
 def main(
     ctx: click.Context,
@@ -83,6 +101,9 @@ def main(
     skip_news: bool,
     workspace_directory: str,
     install_plugin_deps: bool,
+    ai_name: Optional[str],
+    ai_role: Optional[str],
+    ai_goal: tuple[str],
 ) -> None:
     """
     Welcome to AutoGPT an experimental open-source application showcasing the capabilities of the GPT-4 pushing the boundaries of AI.
@@ -109,6 +130,9 @@ def main(
             skip_news,
             workspace_directory,
             install_plugin_deps,
+            ai_name,
+            ai_role,
+            ai_goal,
         )
 
 

autogpt/command_decorator.py

@@ -20,6 +20,7 @@ def command(
     parameters: dict[str, CommandParameterSpec],
     enabled: bool | Callable[[Config], bool] = True,
     disabled_reason: Optional[str] = None,
+    aliases: list[str] = [],
 ) -> Callable[..., Any]:
     """The command decorator is used to create Command objects from ordinary functions."""
 
@@ -40,6 +41,7 @@ def decorator(func: Callable[..., Any]) -> Command:
             parameters=typed_parameters,
             enabled=enabled,
             disabled_reason=disabled_reason,
+            aliases=aliases,
         )
 
         @functools.wraps(func)

autogpt/commands/file_operations.py

@@ -189,6 +189,7 @@ def ingest_file(
             "required": True,
         },
     },
+    aliases=["write_file", "create_file"],
 )
 def write_to_file(filename: str, text: str, agent: Agent) -> str:
     """Write text to a file