siada-cli 1.0.0__tar.gz → 1.2.0__tar.gz

{siada_cli-1.0.0 → siada_cli-1.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: siada-cli
-Version: 1.0.0
+Version: 1.2.0
 Summary: Siada CLI is a Ai Pair Programming Tool in terminal
 License: Apache-2.0
 Requires-Python: >=3.12,<4.0
@@ -12,31 +12,35 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Software Development
+Requires-Dist: aiohttp (>=3.8.0,<4.0.0)
 Requires-Dist: binaryornot (==0.4.4)
 Requires-Dist: click (>=8.2.1,<9.0.0)
 Requires-Dist: configargparse (>=1.7,<2.0.0)
 Requires-Dist: diskcache (==5.6.3)
 Requires-Dist: fastapi (>=0.110.0,<1.0.0)
+Requires-Dist: gitpython (>=3.1.0,<4.0.0)
 Requires-Dist: grep-ast (>=0.9.0,<1.0.0)
 Requires-Dist: litellm (>=1.72.6,<2.0.0)
-Requires-Dist: openai (>=1.99.9,<2.0.0)
-Requires-Dist: openai-agents (==0.2.5)
+Requires-Dist: mcp (>=1.0.0,<2.0.0)
+Requires-Dist: networkx (>=3.5.0,<4.0.0)
+Requires-Dist: openai (>=1.102.0)
+Requires-Dist: openai-agents (==0.2.9)
 Requires-Dist: openhands-aci (==0.3.0)
 Requires-Dist: partial-json-parser (>=0.2.1.1.post6,<1.0.0)
 Requires-Dist: patch-ng (>=1.18.1,<2.0.0)
 Requires-Dist: pexpect (>=4.9.0,<5.0.0)
+Requires-Dist: playwright (>=1.40.0,<2.0.0)
 Requires-Dist: prompt-toolkit (>=3.0.51,<4.0.0)
 Requires-Dist: psutil (>=7.0.0,<8.0.0)
 Requires-Dist: pydantic (>=2.6.0,<3.0.0)
 Requires-Dist: pydantic-settings (>=2.2.0,<3.0.0)
 Requires-Dist: pygments (>=2.19.1,<3.0.0)
-Requires-Dist: rich (>=13.0.0,<14.0.0)
+Requires-Dist: rich (>=13.0.0,<=14.1.0)
 Requires-Dist: shtab (>=1.7,<2.0.0)
 Requires-Dist: termcolor (>=3.1.0,<4.0.0)
 Requires-Dist: tqdm (>=4.67.1,<5.0.0)
 Requires-Dist: trafilatura (>=2.0.0,<3.0.0)
 Requires-Dist: uvicorn (>=0.27.0,<1.0.0)
-Project-URL: Homepage, https://github.com/liauto-siada/siada-cli
 Description-Content-Type: text/markdown
 
 # Siada CLI
@@ -204,6 +208,77 @@ export OPENROUTER_API_KEY="your_openrouter_key"
 unset SIADA_MODEL
 ```
 
+### Checkpoints Configuration
+
+Siada CLI provides checkpoint tracking functionality to automatically save session states and enable recovery from previous points in your development workflow.
+
+**What are Checkpoints?**
+- Automatic snapshots of your session state after significant tool operations
+- Includes conversation history, modified files, and git state
+- Enables rollback to previous states and comparison between different points in time
+
+**Enable Checkpoint Tracking:**
+
+   **Method 1: Command Line Parameter**
+   ```bash
+   # Enable checkpoints when starting CLI
+   siada-cli --checkpointing
+   ```
+
+   **Method 2: Environment Variable**
+   ```bash
+   # Enable checkpoints globally
+   export SIADA_CHECKPOINTING=true
+   siada-cli --agent coder
+   ```
+
+   **Method 3: Configuration File**
+   
+   edit `~/.siada-cli/conf.yaml`:
+   ```yaml
+   checkpoint_config:
+     enable: true
+   ```
+
+**Checkpoint Usage:**
+- Checkpoints are automatically created after tool operations like file edits and command executions
+- Use `/restore <checkpoint_file>` to restore to a previous state
+- Use `/undo <checkpoint_file>` to undo changes made by a checkpoint, restoring to the state before the checkpoint was created
+- Use `/compare <checkpoint_file>` to see differences between current state and a checkpoint
+
+**Storage Location:**
+- location: `~/.siada-cli/data/tmp/{project_hash}/checkpoints/session_id/`
+
+### MCP Configuration
+
+Siada CLI integrates MCP (Model Context Protocol) service to provide extended tools and resources for AI agents.
+
+**MCP Configuration File `~/.siada-cli/mcp_config.json`**
+
+   Parameter descriptions:
+   - `enabled`: Controls global/individual MCP server switches
+   - `type`: Connection type (`stdio`/`http`/`sse`)
+
+   Configuration file example:
+   ```json
+   {
+      "enabled": true,
+      "mcpServers": {
+         "filesystem": {
+            "enabled": true,
+            "type": "stdio",
+            "command": "npx",
+            "args": ["-y", "@modelcontextprotocolserver-filesystem", "./"]
+         }
+      }
+   }
+   ```
+
+**MCP Slash Commands**
+
+   - `/mcp-server`: List all MCP servers
+   - `/mcp-list`: List all MCP servers and their available tools
+
 ## Usage Modes
 
 Siada CLI supports two usage modes to meet different usage scenarios:
@@ -276,6 +351,15 @@ siada-cli --verbose
 # List all available models
 siada-cli --list-models
 siada-cli --models
+
+# Enable checkpoint tracking (for session recovery)
+siada-cli --checkpointing
+
+## Version Check and Update
+siada-cli --just-check-update  # Check version only, without executing update
+siada-cli --upgrade            # Upgrade to the latest version immediately
+siada-cli --check-update       # Check and prompt for updates on startup (enabled by default)
+
 ```
 
 ## Slash Commands
@@ -285,7 +369,17 @@ In the CLI, you can use slash commands for additional functionality:
 - `/shell` - Switch to shell mode to execute system commands (type `exit` or `quit` to exit shell mode)
 - `/models` - List available AI models
 - `/run <command>` or `!<command>` - Execute shell commands
-- `/editor` - Open editor for multiline input
+- `/editor` or `/edit` - Open editor for multiline input
+- `/multiline-mode` - Toggle multiline mode (changes behavior of Enter and Meta+Enter keys)
+  - Enter key for line break
+  - Meta+Enter key to end multiline mode and send content to the model
+- `/init [--force]` - Analyze the project and create a tailored siada.md file
+- `/restore <checkpoint_file>` - Restore session state from a checkpoint file (requires checkpoint tracking enabled)
+- `/undo <checkpoint_file>` - Undo changes made by a checkpoint, restoring to the state before the checkpoint was created (requires checkpoint tracking enabled)
+- `/compare <checkpoint_file>` - Compare current state with a checkpoint file to see differences
+- `/memory-refresh` - Refresh user memory content from siada.md file
+- `/memory-status` - Display current user memory status (file info, size, loaded status)
+- `/status` - Display current session status (model, agent, session ID, and workspace)
 - `/exit` or `/quit` - Exit the application
 
 ### Shell Mode Usage Guide
@@ -475,3 +569,4 @@ See [disclaimers.md](./disclaimers.md)
 <div align="center">
 Built with ❤️ by Li Auto Code Intelligence Team and the open source community
 </div>
+

{siada_cli-1.0.0 → siada_cli-1.2.0}/README.md

@@ -163,6 +163,77 @@ export OPENROUTER_API_KEY="your_openrouter_key"
 unset SIADA_MODEL
 ```
 
+### Checkpoints Configuration
+
+Siada CLI provides checkpoint tracking functionality to automatically save session states and enable recovery from previous points in your development workflow.
+
+**What are Checkpoints?**
+- Automatic snapshots of your session state after significant tool operations
+- Includes conversation history, modified files, and git state
+- Enables rollback to previous states and comparison between different points in time
+
+**Enable Checkpoint Tracking:**
+
+   **Method 1: Command Line Parameter**
+   ```bash
+   # Enable checkpoints when starting CLI
+   siada-cli --checkpointing
+   ```
+
+   **Method 2: Environment Variable**
+   ```bash
+   # Enable checkpoints globally
+   export SIADA_CHECKPOINTING=true
+   siada-cli --agent coder
+   ```
+
+   **Method 3: Configuration File**
+   
+   edit `~/.siada-cli/conf.yaml`:
+   ```yaml
+   checkpoint_config:
+     enable: true
+   ```
+
+**Checkpoint Usage:**
+- Checkpoints are automatically created after tool operations like file edits and command executions
+- Use `/restore <checkpoint_file>` to restore to a previous state
+- Use `/undo <checkpoint_file>` to undo changes made by a checkpoint, restoring to the state before the checkpoint was created
+- Use `/compare <checkpoint_file>` to see differences between current state and a checkpoint
+
+**Storage Location:**
+- location: `~/.siada-cli/data/tmp/{project_hash}/checkpoints/session_id/`
+
+### MCP Configuration
+
+Siada CLI integrates MCP (Model Context Protocol) service to provide extended tools and resources for AI agents.
+
+**MCP Configuration File `~/.siada-cli/mcp_config.json`**
+
+   Parameter descriptions:
+   - `enabled`: Controls global/individual MCP server switches
+   - `type`: Connection type (`stdio`/`http`/`sse`)
+
+   Configuration file example:
+   ```json
+   {
+      "enabled": true,
+      "mcpServers": {
+         "filesystem": {
+            "enabled": true,
+            "type": "stdio",
+            "command": "npx",
+            "args": ["-y", "@modelcontextprotocolserver-filesystem", "./"]
+         }
+      }
+   }
+   ```
+
+**MCP Slash Commands**
+
+   - `/mcp-server`: List all MCP servers
+   - `/mcp-list`: List all MCP servers and their available tools
+
 ## Usage Modes
 
 Siada CLI supports two usage modes to meet different usage scenarios:
@@ -235,6 +306,15 @@ siada-cli --verbose
 # List all available models
 siada-cli --list-models
 siada-cli --models
+
+# Enable checkpoint tracking (for session recovery)
+siada-cli --checkpointing
+
+## Version Check and Update
+siada-cli --just-check-update  # Check version only, without executing update
+siada-cli --upgrade            # Upgrade to the latest version immediately
+siada-cli --check-update       # Check and prompt for updates on startup (enabled by default)
+
 ```
 
 ## Slash Commands
@@ -244,7 +324,17 @@ In the CLI, you can use slash commands for additional functionality:
 - `/shell` - Switch to shell mode to execute system commands (type `exit` or `quit` to exit shell mode)
 - `/models` - List available AI models
 - `/run <command>` or `!<command>` - Execute shell commands
-- `/editor` - Open editor for multiline input
+- `/editor` or `/edit` - Open editor for multiline input
+- `/multiline-mode` - Toggle multiline mode (changes behavior of Enter and Meta+Enter keys)
+  - Enter key for line break
+  - Meta+Enter key to end multiline mode and send content to the model
+- `/init [--force]` - Analyze the project and create a tailored siada.md file
+- `/restore <checkpoint_file>` - Restore session state from a checkpoint file (requires checkpoint tracking enabled)
+- `/undo <checkpoint_file>` - Undo changes made by a checkpoint, restoring to the state before the checkpoint was created (requires checkpoint tracking enabled)
+- `/compare <checkpoint_file>` - Compare current state with a checkpoint file to see differences
+- `/memory-refresh` - Refresh user memory content from siada.md file
+- `/memory-status` - Display current user memory status (file info, size, loaded status)
+- `/status` - Display current session status (model, agent, session ID, and workspace)
 - `/exit` or `/quit` - Exit the application
 
 ### Shell Mode Usage Guide
@@ -433,4 +523,4 @@ See [disclaimers.md](./disclaimers.md)
 ----
 <div align="center">
 Built with ❤️ by Li Auto Code Intelligence Team and the open source community
-</div>
+</div>

{siada_cli-1.0.0 → siada_cli-1.2.0}/agent_config.yaml

@@ -7,22 +7,43 @@ agents:
     class: "siada.agent_hub.coder.bug_fix_agent.BugFixAgent"
     description: "Specialized agent for code bug fixing"
     enabled: true
+    supported_modes: "non_interactive"
   
   # Code Development Agent (not implemented yet)
   coder:
     class: "siada.agent_hub.coder.code_gen_agent.CodeGenAgent"
     description: "General-purpose code development agent"
     enabled: true
+    supported_modes: "all"
   
   # Frontend Generation Agent
   fegen:
     class: "siada.agent_hub.coder.fe_gen_agent.FeGenAgent"
     description: "Frontend code generation agent"
     enabled: false
+    supported_modes: "non_interactive"
+
+  browser:
+    class: "siada.agent_hub.coder.browser_agent.BrowserAgent"
+    description: "Browser operate agent"
+    enabled: true
+    supported_modes: "interactive"
+
+  bugreproduce:
+    class: "siada.agent_hub.coder.bug_reproduce_agent.BugReproduceAgent"
+    description: "Bug reproduction agent"
+    enabled: false
+    supported_modes: "non_interactive"
+
+  test:
+    class: "siada.agent_hub.coder.test_agent.TestAgent"
+    description: "Test agent"
+    enabled: false
+    supported_modes: "non_interactive"
 
 # LLM Configuration
 llm_config:
-  provider: "openrouter"                    # Optional：“openrouter”
+  provider: "li"                    # Optional： ”li“ ， “openrouter”
   model_name: "claude-sonnet-4"
   repo_map_tokens: 8192                     # repo map max token
   repo_map_mul_no_files: 16                 # Multiple without chat files

{siada_cli-1.0.0 → siada_cli-1.2.0}/pyproject.toml

@@ -1,12 +1,12 @@
 [tool.poetry]
 name = "siada-cli"
-version = "1.0.0"
+version = "1.2.0"
 description = "Siada CLI is a Ai Pair Programming Tool in terminal"
 packages = [{include = "siada"}, {include = "agent_config.yaml"}]
 include = ["agent_config.yaml"]
 license = "Apache-2.0"
 readme = "README.md"
-
+ 
 classifiers = [
     "Environment :: Console",
     "Intended Audience :: Developers",
@@ -16,9 +16,6 @@ classifiers = [
     "Topic :: Software Development",
 ]
 
-[tool.poetry.urls]
-Homepage = "https://github.com/liauto-siada/siada-cli"
-
 [tool.poetry.scripts]
 siada-cli = "siada.entrypoint.siadahub:main"
 [tool.poetry.dependencies]
@@ -27,8 +24,8 @@ fastapi = ">=0.110.0,<1.0.0"
 uvicorn = ">=0.27.0,<1.0.0"
 pydantic = ">=2.6.0,<3.0.0"
 pydantic-settings = ">=2.2.0,<3.0.0"
-openai-agents = "0.2.5"
-openai = ">=1.99.9,<2.0.0"
+openai-agents = "0.2.9"
+openai = ">=1.102.0"
 litellm = ">=1.72.6,<2.0.0"
 openhands-aci = "0.3.0"
 termcolor = ">=3.1.0,<4.0.0"
@@ -36,7 +33,7 @@ psutil = ">=7.0.0,<8.0.0"
 pexpect = ">=4.9.0,<5.0.0"
 binaryornot = "0.4.4"
 diskcache = "5.6.3"
-rich = ">=13.0.0,<14.0.0"
+rich = ">=13.0.0,<=14.1.0"
 tqdm = ">=4.67.1,<5.0.0"
 pygments = ">=2.19.1,<3.0.0"
 grep-ast = ">=0.9.0,<1.0.0"
@@ -47,6 +44,12 @@ patch-ng = ">=1.18.1,<2.0.0"
 trafilatura = ">=2.0.0,<3.0.0"
 prompt-toolkit = ">=3.0.51,<4.0.0"
 partial-json-parser = ">=0.2.1.1.post6,<1.0.0"
+networkx = ">=3.5.0,<4.0.0"
+playwright = ">=1.40.0,<2.0.0"
+gitpython = ">=3.1.0,<4.0.0"
+mcp = ">=1.0.0,<2.0.0"
+aiohttp = ">=3.8.0,<4.0.0"
+
 
 [tool.poetry.group.dev.dependencies]
 pytest = ">=8.0.0,<9.0.0"

siada_cli-1.2.0/siada/agent_hub/__init__.py

@@ -0,0 +1,10 @@
+"""
+Agent Module
+
+Provides predefined Agent instances and initialization functionality
+"""
+from typing import Dict, Any
+
+
+# Store created Agents, key is agent name, value is agent instance
+agent_map: Dict[str, Any] = {}

siada_cli-1.2.0/siada/agent_hub/coder/browser_agent.py

@@ -0,0 +1,60 @@
+from agents import RunContextWrapper
+
+from siada.foundation.code_agent_context import CodeAgentContext
+from siada.agent_hub.coder.code_gen_agent import CodeGenAgent
+from siada.agent_hub.coder.prompt import fe_gen_prompt
+from siada.tools.browser.browser_action_tool import browser_operate
+from siada.tools.browser.browsergym_action_tool import browser_operate_by_gym
+from siada.tools.coder.file_operator import edit
+from siada.tools.coder.file_search import regex_search_files
+from siada.tools.coder.run_cmd import run_cmd
+
+
+class BrowserAgent(CodeGenAgent):
+
+    def __init__(self, *args, **kwargs):
+
+        super().__init__(
+            name="BrowserAgent",
+            #tools=[edit, regex_search_files, run_cmd, browser_operate],
+            tools=[browser_operate_by_gym],
+            *args,
+            **kwargs
+        )
+
+    async def get_system_prompt(self, run_context: RunContextWrapper[CodeAgentContext]) -> str | None:
+        # 使用专门的浏览器操作提示词，而不是前端生成提示词
+        instructions = """
+You are a Browser Operation Agent specialized in web automation tasks.
+
+Your primary responsibility is to perform browser operations according to user instructions using the browser_operate tool.
+
+IMPORTANT GUIDELINES:
+1. **Screenshot Analysis**: Always carefully analyze the screenshot after each action to understand the current page state
+2. **Coordinate Accuracy**: When clicking elements, examine the screenshot to determine precise coordinates
+3. **Element Identification**: Look for visual cues like buttons, input fields, links, and other interactive elements
+4. **Wait for Loading**: Allow sufficient time for pages to load before taking actions
+5. **Error Handling**: If an action doesn't produce the expected result, analyze the screenshot and try alternative approaches
+
+BROWSER OPERATION WORKFLOW:
+1. Launch browser with the target URL
+2. Wait for page to fully load
+3. Analyze the screenshot to identify target elements
+4. Perform actions (click, type, scroll) based on visual analysis
+5. Verify results by examining subsequent screenshots
+6. Close browser when task is complete
+
+COORDINATE SELECTION TIPS:
+- Click in the center of buttons and input fields
+- For text input, click on the input field first, then type
+- For search operations, locate the search button visually
+- Use scroll actions if elements are not visible
+
+Remember: You can see the page through screenshots, so use this visual information to make accurate decisions about where to click and what actions to take.
+        """
+        return instructions
+
+    async def get_context(self) -> CodeAgentContext:
+        current_working_dir = "/Users/yunan/code/test/fe_gen"
+        context = CodeAgentContext(root_dir=current_working_dir)
+        return context