Skip to content

🌟 The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming

update: 软件公司重构

liushaojief9b0262c5556
diff --git a/metagpt/actions/search_enhanced_qa.py b/metagpt/actions/search_enhanced_qa.py
index 1427f9b19..b1810f06a 100644
--- a/metagpt/actions/search_enhanced_qa.py
+++ b/metagpt/actions/search_enhanced_qa.py
@@ -209,7 +209,7 @@ class SearchEnhancedQA(Action):
list[str]: Summaries of relevant web content.
"""
- relevant_urls = await self._collect_relevant_links(query)
+ relevant_urls = await self.collect_relevant_links(query)
await self._reporter.async_report({"type": "search", "stage": "searching", "urls": relevant_urls})
if not relevant_urls:
logger.warning(f"No relevant URLs found for query: {query}")
@@ -226,7 +226,7 @@ class SearchEnhancedQA(Action):
return citations
- async def _collect_relevant_links(self, query: str) -> list[str]:
+ async def collect_relevant_links(self, query: str) -> list[str]:
"""Search and rank URLs relevant to the query.
Args:
diff --git a/metagpt/prompts/product_manager.py b/metagpt/prompts/product_manager.py
new file mode 100644
index 000000000..0fc7c3557
--- /dev/null
+++ b/metagpt/prompts/product_manager.py
@@ -0,0 +1,175 @@
+from metagpt.prompts.di.role_zero import ROLE_INSTRUCTION
+
+EXTRA_INSTRUCTION = """
+You are a product manager AI assistant specializing in product requirement documentation and market research analysis.
+Your work focuses on the analysis of problems and data.
+You should always output a document.
+
+## Core Tools
+1. Editor: For the creation and modification of `PRD/Research Report` documents.
+2. SearchEnhancedQA: The specified tool for collecting information from the internet MUST BE USED for searching.
+3. Browser: Access the search results provided by the SearchEnhancedQA tool using the "goto" method.
+
+## Mode 1: PRD Creation
+Triggered by software/product requests or feature enhancements, ending with the output of a complete PRD.
+
+### Required Fields
+1. Language & Project Info
+ - Language: Match user's language
+ - Programming Language: Default to native HTML if unspecified
+ - Project Name: Use snake_case format
+ - Restate the original requirements
+
+2. Product Definition(**IMPORTANT** )
+ - Product Goals: 3 clear, orthogonal goals
+ - User Stories: 3-5 scenarios in "As a [role], I want [feature] so that [benefit]" format
+ - Competitive Analysis: 5-7 products with pros/cons
+ - Competitive Quadrant Chart(Required): Using Mermaid
+
+3. Technical Specifications
+ - Requirements Analysis: Comprehensive overview of technical needs
+ - Requirements Pool: List with P0/P1/P2 priorities
+ - UI Design Draft: Basic layout and functionality
+ - Open Questions: Unclear aspects needing clarification
+
+#### Mermaid Diagram Rules
+1. Use mermaid quadrantChart syntax. Distribute scores evenly between 0 and 1
+2. Example:
+```mermaid
+quadrantChart
+ title "Reach and engagement of campaigns"
+ x-axis "Low Reach" --> "High Reach"
+ y-axis "Low Engagement" --> "High Engagement"
+ quadrant-1 "We should expand"
+ quadrant-2 "Need to promote"
+ quadrant-3 "Re-evaluate"
+ quadrant-4 "May be improved"
+ "Campaign A": [0.3, 0.6]
+ "Campaign B": [0.45, 0.23]
+ "Campaign C": [0.57, 0.69]
+ "Campaign D": [0.78, 0.34]
+ "Campaign E": [0.40, 0.34]
+ "Campaign F": [0.35, 0.78]
+ "Our Target Product": [0.5, 0.6]
+```
+
+### PRD Document Guidelines
+- Use clear requirement language (Must/Should/May)
+- Include measurable criteria
+- Prioritize clearly (P0: Must-have, P1: Should-have, P2: Nice-to-have)
+- Support with diagrams and charts
+- Focus on user value and business goals
+
+## Mode 2: Market Research
+Triggered by market analysis or competitor research requests, ending with the output of a complete report document.
+
+### **IMPORTANT** Information Collection Requirements
+
+Must follow this strict information gathering process:
+1. Keyword Generation Rules:
+ - Infer 3 distinct keyword groups on user needs.
+ - Each group must be a space-separated phrase containing:
+ * Target industry/product name (REQUIRED)
+ * Specific aspect or metric
+ * Time frame or geographic scope when relevant
+
+ Example format:
+ - Group 1: "electric vehicles market size forecast 2024"
+ - Group 2: "electric vehicles manufacturing costs analysis"
+ - Group 3: "electric vehicles consumer preferences survey"
+
+2. Search Process:
+ - For each keyword:
+ * Use SearchEnhancedQA TOOL (SearchEnhancedQA.collect_relevant_links) collect top 3 search results
+ * Remove duplicate URLs
+
+3. Information Analysis:
+ - Must read and analyze EACH unique source individually
+ - Synthesize information across all sources
+ - Cross-reference and verify key data points
+ - Identify critical insights and trends
+
+4. Quality Control:
+ - Verify data consistency across sources
+ - Fill information gaps with targeted additional research
+ - Ensure balanced perspective from multiple sources
+
+
+### Report Structure
+1. Summary: Key findings and recommendations
+2. Industry Overview: Market size, trends, and structure
+3. Market Analysis: Segments, growth drivers, and challenges
+4. Competitor Landscape: Key players and positioning
+5. Target Audience Analysis: User segments and needs
+6. Pricing Analysis: Market rates and strategies
+7. Key Findings: Major insights and opportunities
+8. Strategic Recommendations: Action items
+9. Appendices: Supporting data
+
+
+### Final Report Requirements
+1. Report must be entirely focused on insights and analysis:
+ - No mention of research methodology
+ - No source tracking or process documentation
+ - Present only validated findings and conclusions
+
+2. Professional Format:
+ - Clear section hierarchy
+ - Rich subsection content
+ - Evidence-based analysis
+ - Data visualization where appropriate
+
+3. Content Depth Requirements:
+ Executive Summary (500+ words):
+ - Key Market Metrics
+ - Critical Findings
+ - Strategic Recommendations
+
+ Industry Overview (800+ words):
+ - Market Size and Growth
+ - Industry Value Chain
+ - Regulatory Environment
+ - Technology Trends
+
+4. Quality Standards:
+ - Every main section must have 3+ detailed subsections
+ - Each subsection requires 200-300 words minimum
+ - Include specific examples and data points
+ - Support all major claims with market evidence
+
+### Research Guidelines
+- Base all analysis on collected data
+- Include quantitative and qualitative insights
+- Support claims with evidence
+- Maintain professional formatting
+- Use visuals to support key points
+
+## Document Standards
+1. Format
+ - Clear heading hierarchy
+ - Consistent markdown formatting
+ - Numbered sections
+ - Professional graphics
+ - Output charts using Mermaid syntax
+
+2. Content
+ - Objective analysis
+ - Actionable insights
+ - Clear recommendations
+ - Supporting evidence
+
+3. Quality Checks
+ - Verify data accuracy
+ - Cross-reference sources
+ - Ensure completeness
+ - Review clarity
+
+Remember:
+- Always start with thorough requirements analysis
+- Use appropriate tools for each task
+- Keep recommendations actionable
+- Consider all stakeholder perspectives
+- Maintain professional standards throughout
+"""
+
+PRODUCT_MANAGER_INSTRUCTION = ROLE_INSTRUCTION + EXTRA_INSTRUCTION.strip()
diff --git a/metagpt/roles/product_manager.py b/metagpt/roles/product_manager.py
index f2a47affd..e91ad86e7 100644
--- a/metagpt/roles/product_manager.py
+++ b/metagpt/roles/product_manager.py
@@ -4,16 +4,14 @@
@Time : 2023/5/11 14:43
@Author : alexanderwu
@File : product_manager.py
-@Modified By: mashenquan, 2023/11/27. Add `PrepareDocuments` action according to Section 2.2.3.5.1 of RFC 135.
+@Modified By: liushaojie, 2024/10/17.
"""
-
-from metagpt.actions import UserRequirement, WritePRD
-from metagpt.actions.prepare_documents import PrepareDocuments
-from metagpt.actions.requirement_analysis.requirement.pic2txt import Pic2Txt
+from metagpt.actions.search_enhanced_qa import SearchEnhancedQA
+from metagpt.prompts.product_manager import PRODUCT_MANAGER_INSTRUCTION
from metagpt.roles.di.role_zero import RoleZero
-from metagpt.roles.role import RoleReactMode
-from metagpt.utils.common import any_to_name, any_to_str, tool2name
-from metagpt.utils.git_repository import GitRepository
+from metagpt.tools.libs.browser import Browser
+from metagpt.tools.libs.editor import Editor
+from metagpt.utils.common import tool2name
class ProductManager(RoleZero):
@@ -29,38 +27,18 @@ class ProductManager(RoleZero):
name: str = "Alice"
profile: str = "Product Manager"
- goal: str = "efficiently create a successful product that meets market demands and user expectations. Create a Product Requirement Document."
+ goal: str = "Create a Product Requirement Document or market research/competitive product research."
constraints: str = "utilize the same language as the user requirements for seamless communication"
- todo_action: str = any_to_name(WritePRD)
-
- instruction: str = """Use WritePRD tool to write PRD if a PRD is required, users may asks for a software without mentioning PRD, but you should output the PRD of that software; Use `Pic2Txt` tool to write out an intact textual user requirements if an intact textual user requiremnt is required given some images alongside the contextual textual descriptions"""
- max_react_loop: int = 1 # FIXME: Read and edit files requires more steps, consider later
- tools: list[str] = ["RoleZero", "WritePRD", Pic2Txt.__name__]
+ instruction: str = PRODUCT_MANAGER_INSTRUCTION
+ max_react_loop: int = 50
+ tools: list[str] = ["RoleZero", Browser.__name__, Editor.__name__, SearchEnhancedQA.__name__]
def __init__(self, **kwargs) -> None:
super().__init__(**kwargs)
- # NOTE: The following init setting will only be effective when self.use_fixed_sop is changed to True
self.enable_memory = False
- self.set_actions([PrepareDocuments(send_to=any_to_str(self)), WritePRD])
- if self.use_fixed_sop:
- self._watch([UserRequirement, PrepareDocuments])
- self.rc.react_mode = RoleReactMode.BY_ORDER
def _update_tool_execution(self):
- wp = WritePRD()
- self.tool_execution_map.update(tool2name(WritePRD, ["run"], wp.run))
- pic2txt = Pic2Txt()
- self.tool_execution_map.update(tool2name(Pic2Txt, ["run"], pic2txt.run))
-
- async def _think(self) -> bool:
- """Decide what to do"""
- if not self.use_fixed_sop:
- return await super()._think()
-
- if GitRepository.is_git_dir(self.config.project_path) and not self.config.git_reinit:
- self._set_state(1)
- else:
- self._set_state(0)
- self.config.git_reinit = False
- self.todo_action = any_to_name(WritePRD)
- return bool(self.rc.todo)
+ se_qa = SearchEnhancedQA()
+ self.tool_execution_map.update(
+ tool2name(SearchEnhancedQA, ["collect_relevant_links"], se_qa.collect_relevant_links)
+ )
diff --git a/metagpt/schema.py b/metagpt/schema.py
index 5224eaf14..24ca5b14f 100644
--- a/metagpt/schema.py
+++ b/metagpt/schema.py
@@ -685,7 +685,17 @@ class Plan(BaseModel):
def append_task(
self, task_id: str, dependent_task_ids: list[str], instruction: str, assignee: str, task_type: str = ""
):
- """Append a new task with task_id (number) to the end of existing task sequences. If dependent_task_ids is not empty, the task will depend on the tasks with the ids in the list. Note that the assignee should be the 'name' of the role."""
+ """
+ Append a new task with task_id (number) to the end of existing task sequences.
+ If dependent_task_ids is not empty, the task will depend on the tasks with the ids in the list.
+ Note that the assignee should be the 'name' of the role.
+ Args:
+ task_id (str): The task id to be appended to the existing task sequence
+ dependent_task_ids (list[str]): The task ids that the new task depends on
+ instruction (str): The instruction of the new task
+ assignee (str): The assignee of the new task
+ task_type (str): The type of the new task, default is empty string
+ """
new_task = Task(
task_id=task_id,
dependent_task_ids=dependent_task_ids,
diff --git a/tests/metagpt/roles/di/run_product_manager.py b/tests/metagpt/roles/di/run_product_manager.py
index 3ab1e9bab..42aa7840a 100644
--- a/tests/metagpt/roles/di/run_product_manager.py
+++ b/tests/metagpt/roles/di/run_product_manager.py
@@ -1,13 +1,72 @@
import asyncio
+import sys
-from metagpt.roles.product_manager import ProductManager
+from metagpt.logs import logger
+from metagpt.roles import ProductManager
WRITE_2048 = """Write a PRD for a cli 2048 game"""
-REWRITE_2048 = """Rewrite the prd at /Users/gary/Files/temp/workspace/2048_game/docs/prd.json, add a web UI"""
+# REWRITE_2048 = """Rewrite the prd at /Users/gary/Files/temp/workspace/2048_game/docs/prd.json, add a web UI"""
CASUAL_CHAT = """What's your name?"""
+CASE1_GREEDY_SNAKE = "设计一个贪吃蛇游戏,root path: '/Users/seeker/Projects/sdfz/mg/mgx_ops/workspace/temp'"
+CASE2_SMART_HOME = "搜索并分析米家、华为智能家居和海尔智家在智能家居市场中的功能、用户需求和市场定位"
+CASE3_BEST_SELLING_REFRIGERATOR = "调研当前市场上最畅销的智能冰箱的五个关键特性"
+
+OLD_PRD = """
+Language
+en_us
+
+Programming Language
+N/A
+
+Original Requirements
+Write a PRD based on the current music streaming service.
+
+Project Name
+music_streaming_service
+
+Product Goals
+Enhance user experience with seamless music streaming
+Improve accessibility and responsiveness across devices
+Expand music library and personalized recommendations
+User Stories
+As a user, I want to easily search and find my favorite songs and artists.
+As a user, I want to create and manage my own playlists.
+As a user, I want to receive personalized music recommendations based on my listening history.
+As a user, I want to stream music without interruptions or buffering.
+As a user, I want to access the service on both desktop and mobile devices.
+Competitive Analysis
+Spotify: Extensive music library, strong personalized recommendations, and cross-platform availability.
+Apple Music: High-quality audio, exclusive content, and seamless integration with Apple devices.
+Amazon Music: Large music catalog, integration with Amazon Echo devices, and competitive pricing.
+YouTube Music: Vast collection of music videos, user-generated content, and strong search capabilities.
+Tidal: High-fidelity sound quality, exclusive releases, and artist-centric approach.
+Competitive Quadrant Chart
+quadrantChart title "Feature Richness vs. User Satisfaction" x-axis "Low Feature Richness" --> "High Feature Richness" y-axis "Low User Satisfaction" --> "High User Satisfaction" quadrant-1 "Market Leaders" quadrant-2 "Potential Growth" quadrant-3 "Needs Improvement" quadrant-4 "Niche Players" "Spotify": [0.9, 0.85] "Apple Music": [0.85, 0.8] "Amazon Music": [0.75, 0.7] "YouTube Music": [0.8, 0.75] "Tidal": [0.7, 0.65] "Our Target Product": [0.8, 0.8]
+
+Requirement Analysis
+The current music streaming service needs to focus on enhancing user experience by providing seamless streaming, improving accessibility, and expanding the music library. Personalized recommendations and cross-platform availability are crucial for user retention.
+
+Requirement Pool
+['P0', 'Implement a robust search functionality to find songs and artists easily.']
+['P0', 'Develop a feature for users to create and manage playlists.']
+['P1', 'Enhance the recommendation algorithm for personalized music suggestions.']
+['P1', 'Optimize the streaming service to minimize interruptions and buffering.']
+['P2', 'Ensure the service is fully responsive and accessible on both desktop and mobile devices.']
+UI Design draft
+The UI should be clean and intuitive, with a prominent search bar, easy-to-navigate menus for playlists and recommendations, and a responsive design that adapts to different screen sizes. The player controls should be easily accessible, and the overall aesthetic should be modern and visually appealing.
+
+Anything UNCLEAR
+Currently, all aspects of the project are clear.
+"""
+CASE4_MUSIC_STREAMING_MEDIA = f"""We have received feedback from users regarding the current music streaming service, stating that they need better personalized recommendations. Please readjust the content of PRD {OLD_PRD} based on these feedback."""
+
+CASE5_SMART_BIG_SCREEN = """分析2024年上半年中国家庭智能大屏行业的发展情况并输出市场分析报告"""
+
+# CASE6 = "我想要生产一个电子烟产品,请帮我完成市场调研分析报告"
+
async def main(requirement):
product_manager = ProductManager()
@@ -15,4 +74,21 @@ async def main(requirement):
if __name__ == "__main__":
- asyncio.run(main(WRITE_2048))
+ cases = [
+ # CASE1_GREEDY_SNAKE,
+ # CASE2_SMART_HOME,
+ CASE3_BEST_SELLING_REFRIGERATOR,
+ # CASE4_MUSIC_STREAMING_MEDIA,
+ # CASE5_SMART_BIG_SCREEN,
+ # CASE6,
+ # WRITE_2048,
+ ]
+ logger.remove()
+ # logger.add(sys.stderr, level="DEBUG")
+ logger.add(sys.stderr, level="INFO")
+ for case in cases:
+ logger.info(f"user requirement:\n{case}")
+ try:
+ asyncio.run(main(case))
+ except Exception as e:
+ print(e)