zrb 1.9.1__tar.gz → 1.9.3__tar.gz

{zrb-1.9.1 → zrb-1.9.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: zrb
-Version: 1.9.1
+Version: 1.9.3
 Summary: Your Automation Powerhouse
 Home-page: https://github.com/state-alchemists/zrb
 License: AGPL-3.0-or-later
@@ -32,6 +32,7 @@ Requires-Dist: pyjwt (>=2.10.1,<3.0.0)
 Requires-Dist: python-dotenv (>=1.1.1,<2.0.0)
 Requires-Dist: python-jose[cryptography] (>=3.4.0,<4.0.0)
 Requires-Dist: requests (>=2.32.4,<3.0.0)
+Requires-Dist: rich (>=14.0.0,<15.0.0)
 Requires-Dist: tiktoken (>=0.8.0,<0.9.0)
 Requires-Dist: ulid-py (>=1.1.0,<2.0.0)
 Project-URL: Documentation, https://github.com/state-alchemists/zrb
@@ -91,6 +92,7 @@ This guide shows you how to define two simple tasks:
 - And another one to convert that script into a PNG image.
 
 > **Note:** This assume you have an `OPENAI_API_KEY` and a Mermaid CLI installed (i.e., `npm install -g @mermaid-js/mermaid-cli`)
+> If you want to use other LLM Provider (e.g., Deepseek, Ollama, etc.), you can refer to [Zrb LLM Integration guide](https://github.com/state-alchemists/zrb/blob/main/docs/installation-and-configuration/configuration/llm-integration.md)
 
 ## 1. Create Your Task Definition File
 

{zrb-1.9.1 → zrb-1.9.3}/README.md

@@ -51,6 +51,7 @@ This guide shows you how to define two simple tasks:
 - And another one to convert that script into a PNG image.
 
 > **Note:** This assume you have an `OPENAI_API_KEY` and a Mermaid CLI installed (i.e., `npm install -g @mermaid-js/mermaid-cli`)
+> If you want to use other LLM Provider (e.g., Deepseek, Ollama, etc.), you can refer to [Zrb LLM Integration guide](https://github.com/state-alchemists/zrb/blob/main/docs/installation-and-configuration/configuration/llm-integration.md)
 
 ## 1. Create Your Task Definition File
 

{zrb-1.9.1 → zrb-1.9.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "zrb"
-version = "1.9.1"
+version = "1.9.3"
 description = "Your Automation Powerhouse"
 authors = ["Go Frendi Gunawan <gofrendiasgard@gmail.com>"]
 license = "AGPL-3.0-or-later"
@@ -56,6 +56,7 @@ playwright = {version = "^1.53.0", optional = true}
 psutil = "^7.0.0"
 pyjwt = "^2.10.1"
 prompt-toolkit = "^3.0.51"
+rich = "^14.0.0"
 
 [tool.poetry.extras]
 # poetry install -E rag

{zrb-1.9.1 → zrb-1.9.3}/src/zrb/builtin/llm/chat_session.py

@@ -29,7 +29,7 @@ async def read_user_prompt(ctx: AnyContext) -> str:
 
     _show_info(ctx)
     final_result = ""
-    ctx.print(stylize_faint("💬"), plain=True)
+    ctx.print("💬 >>", plain=True)
     ctx.print(ctx.input.message, plain=True)
     ctx.print("", plain=True)
     result = await _trigger_ask_and_wait_for_result(
@@ -49,12 +49,12 @@ async def read_user_prompt(ctx: AnyContext) -> str:
     while True:
         await asyncio.sleep(0.01)
         if not multiline_mode:
-            ctx.print(stylize_faint("💬"), plain=True)
+            ctx.print("💬 >>", plain=True)
         user_input = await user_input_session.prompt_async()
         if not multiline_mode:
             ctx.print("", plain=True)
         # Handle special input
-        if user_input.strip().lower() in ("/bye", "/quit"):
+        if user_input.strip().lower() in ("/bye", "/quit", "/q", "/exit"):
             user_prompt = "\n".join(user_inputs)
             user_inputs = []
             result = await _trigger_ask_and_wait_for_result(ctx, user_prompt)
@@ -108,12 +108,27 @@ async def _trigger_ask_and_wait_for_result(
         return None
     await _trigger_ask(ctx, user_prompt, previous_session_name, start_new)
     result = await _wait_ask_result(ctx)
-    ctx.print(stylize_faint("\n🤖"), plain=True)
-    ctx.print(result, plain=True)
+    md_result = _render_markdown(result) if result is not None else ""
+    ctx.print("\n🤖 >>", plain=True)
+    ctx.print(md_result, plain=True)
     ctx.print("", plain=True)
     return result
 
 
+def _render_markdown(markdown_text: str) -> str:
+    """
+    Renders Markdown to a string, ensuring link URLs are visible.
+    """
+    from rich.console import Console
+    from rich.markdown import Markdown
+
+    console = Console()
+    markdown = Markdown(markdown_text, hyperlinks=False)
+    with console.capture() as capture:
+        console.print(markdown)
+    return capture.get()
+
+
 def get_llm_ask_input_mapping(callback_ctx: AnyContext):
     """
     Generates the input mapping for the LLM ask task from the callback context.

{zrb-1.9.1 → zrb-1.9.3}/src/zrb/builtin/llm/tool/file.py

@@ -415,6 +415,7 @@ def apply_diff(
     replace_content: str,
 ) -> str:
     """Apply a precise search/replace to a file based on line numbers and content.
+    This tool enables you to update certain part of the file efficiently.
     Args:
         path (str): Path to modify. Pass exactly as provided, including '~'.
         start_line (int): The 1-based starting line number of the content to replace.

{zrb-1.9.1 → zrb-1.9.3}/src/zrb/llm_config.py

@@ -9,32 +9,42 @@ if TYPE_CHECKING:
 
 
 DEFAULT_PERSONA = (
-    "You are a helpful and precise expert assistant. Your goal is to follow "
-    "instructions carefully to provide accurate and efficient help. Get "
-    "straight to the point."
+    "You are a helpful and precise expert assistant. Your primary goal is to "
+    "assist the user by proactively using your available tools to find "
+    "information, execute tasks, and solve problems. Always be accurate, "
+    "efficient, and get straight to the point."
 ).strip()
 
 DEFAULT_INTERACTIVE_SYSTEM_PROMPT = (
     "This is an interactive session. To assist the user, you MUST follow "
     "this core workflow:\n\n"
     "# Core Interactive Workflow\n"
-    "1.  **Clarify:** If the user's request is ambiguous, ask clarifying "
-    "questions to ensure you fully understand their goal.\n"
-    "2.  **Plan:** For any non-trivial task, briefly outline your plan to the "
+    "1.  **Understand:** First, understand the user's request. If you "
+    "encounter any concept, term, or topic you are unfamiliar with, you MUST "
+    "use your tools to search for information before proceeding. Do not "
+    "apologize for not knowing something; find the answer.\n"
+    "2.  **Clarify:** If the user's request is ambiguous after your initial "
+    "research, ask clarifying questions to ensure you fully understand their "
+    "goal.\n"
+    "3.  **Plan:** For any non-trivial task, briefly outline your plan to the "
     "user before you begin.\n"
-    "3.  **Execute:** Carry out the plan, using your available tools as "
+    "4.  **Execute:** Carry out the plan, using your available tools as "
     "needed.\n"
-    "4.  **Confirm:** Before performing any significant or irreversible "
+    "5.  **Confirm:** Before performing any significant or irreversible "
     "action (e.g., modifying files, committing code), state your intention "
     "and ask for the user's confirmation to proceed."
 ).strip()
 
 DEFAULT_SYSTEM_PROMPT = (
     "To fulfill this one-shot request, follow this simple process:\n"
-    "1.  **Analyze:** Deconstruct the user's request.\n"
-    "2.  **Execute:** Fulfill the request directly and concisely, using tools "
-    "if necessary.\n"
-    "3.  **Answer:** Provide a clear and accurate answer."
+    "1.  **Analyze and Research:** Deconstruct the user's request. If you "
+    "encounter any concept, term, or topic you are unfamiliar with, you MUST "
+    "use your tools to search for information. Do not state you don't know "
+    "something without trying to find out first.\n"
+    "2.  **Execute:** Fulfill the request directly and concisely, using your "
+    "tools to perform actions and gather any final, necessary information.\n"
+    "3.  **Answer:** Provide a clear, accurate, and well-informed answer "
+    "based on your research and execution."
 ).strip()
 
 DEFAULT_SPECIAL_INSTRUCTION_PROMPT = (
@@ -87,8 +97,10 @@ DEFAULT_SUMMARIZATION_PROMPT = (
     "`Previous Summary`.\n"
     "1.  Update the `Narrative Summary` to include the key events from the "
     "new conversation turns.\n"
-    "2.  Replace the `Recent History` with the last 4 turns of the *new* "
-    "conversation.\n\n"
+    "2.  Replace the `Recent History` with the most recent turns from the *new* "
+    "conversation. It is critical to include the user's very last message "
+    "and assistant's very last response, ensuring the end of the conversation is "
+    "not truncated.\n\n"
     "### 3. Output Specification\n"
     "Your entire output MUST be a single block of text containing the "
     "following two sections in this exact order:\n"

{zrb-1.9.1 → zrb-1.9.3}/src/zrb/task/llm/print_node.py

@@ -18,23 +18,18 @@ async def print_node(print_func: Callable, agent_run: Any, node: Any):
     )
 
     if Agent.is_user_prompt_node(node):
-        # A user prompt node => The user has provided input
-        print_func(stylize_faint(f"    >> UserPromptNode: {node.user_prompt}"))
+        print_func(stylize_faint("🔠 Receiving input..."))
     elif Agent.is_model_request_node(node):
         # A model request node => We can stream tokens from the model's request
-        print_func(
-            stylize_faint("    >> ModelRequestNode: streaming partial request tokens")
-        )
+        print_func(stylize_faint("🧠 Processing..."))
         async with node.stream(agent_run.ctx) as request_stream:
             is_streaming = False
             async for event in request_stream:
-                if isinstance(event, PartStartEvent):
+                if isinstance(event, PartStartEvent) and event.part:
                     if is_streaming:
                         print_func("")
                     print_func(
-                        stylize_faint(
-                            f"    [Request] Starting part {event.index}: {event.part!r}"
-                        ),
+                        stylize_faint(f"    Starting part {event.index}: {event.part}"),
                     )
                     is_streaming = False
                 elif isinstance(event, PartDeltaEvent):
@@ -49,22 +44,18 @@ async def print_node(print_func: Callable, agent_run: Any, node: Any):
                             end="",
                         )
                     is_streaming = True
-                elif isinstance(event, FinalResultEvent):
+                elif isinstance(event, FinalResultEvent) and event.tool_name:
                     if is_streaming:
                         print_func("")
                     print_func(
-                        stylize_faint(f"    [Result] tool_name={event.tool_name}"),
+                        stylize_faint(f"    Result: tool_name={event.tool_name}"),
                     )
                     is_streaming = False
             if is_streaming:
                 print_func("")
     elif Agent.is_call_tools_node(node):
         # A handle-response node => The model returned some data, potentially calls a tool
-        print_func(
-            stylize_faint(
-                "    >> CallToolsNode: streaming partial response & tool usage"
-            )
-        )
+        print_func(stylize_faint("🧰 Calling Tool..."))
         async with node.stream(agent_run.ctx) as handle_stream:
             async for event in handle_stream:
                 if isinstance(event, FunctionToolCallEvent):
@@ -84,16 +75,16 @@ async def print_node(print_func: Callable, agent_run: Any, node: Any):
                         del event.part.args["_dummy"]
                     print_func(
                         stylize_faint(
-                            f"    [Tools] The LLM calls tool={event.part.tool_name!r} with args={event.part.args} (tool_call_id={event.part.tool_call_id!r})"  # noqa
+                            f"    {event.part.tool_call_id} | "
+                            f"Call {event.part.tool_name} {event.part.args}"
                         )
                     )
                 elif isinstance(event, FunctionToolResultEvent):
                     print_func(
                         stylize_faint(
-                            f"    [Tools] Tool call {event.tool_call_id!r} returned => {event.result.content}"  # noqa
+                            f"    {event.tool_call_id} | {event.result.content}"
                         )
                     )
     elif Agent.is_end_node(node):
         # Once an End node is reached, the agent run is complete
-        print_func(stylize_faint("    [End of Response]"))
-        # print_func(stylize_faint(f"{agent_run.result.data}"))
+        print_func(stylize_faint("✅ Completed..."))

{zrb-1.9.1 → zrb-1.9.3}/src/zrb/task/llm_task.py

@@ -356,7 +356,7 @@ class LLMTask(BaseTask):
                     ctx.xcom[xcom_usage_key] = Xcom([])
                 usage = agent_run.result.usage()
                 ctx.xcom[xcom_usage_key].push(usage)
-                ctx.print(stylize_faint(f"    Token: {usage}"), plain=True)
+                ctx.print(stylize_faint(f"💸 Token: {usage}"), plain=True)
                 return agent_run.result.output
             else:
                 ctx.log_warning("Agent run did not produce a result.")