bioguider 0.2.5__tar.gz → 0.2.7__tar.gz

{bioguider-0.2.5 → bioguider-0.2.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: bioguider
-Version: 0.2.5
+Version: 0.2.7
 Summary: An AI-Powered package to help biomedical developers to generate clear documentation
 License: MIT
 Author: Cankun Wang
@@ -28,6 +28,7 @@ Requires-Dist: langchain-experimental (>=0.3.4,<0.4.0)
 Requires-Dist: langchain-google-genai (>=2.1.4,<3.0.0)
 Requires-Dist: langchain-openai (>=0.3.8,<0.4.0)
 Requires-Dist: langgraph (>=0.3.11,<0.4.0)
+Requires-Dist: markdownify (>=1.1.0,<2.0.0)
 Requires-Dist: nanoid (>=2.0.0,<3.0.0)
 Requires-Dist: pydantic (>=2.10.6,<3.0.0)
 Requires-Dist: pydantic-settings (>=2.8.1,<3.0.0)

{bioguider-0.2.5 → bioguider-0.2.7}/bioguider/agents/agent_tools.py

@@ -1,9 +1,11 @@
 import os
 from typing import Callable
+from markdownify import markdownify as md
 from langchain_openai.chat_models.base import BaseChatOpenAI
 from bioguider.database.summarized_file_db import SummarizedFilesDb
 from bioguider.utils.file_utils import get_file_type
 from bioguider.agents.agent_utils import read_directory, read_file, summarize_file
+from bioguider.rag.data_pipeline import count_tokens
 
 class agent_tool:
     def __init__(
@@ -39,7 +41,14 @@ Returns:
             file_path = os.path.join(self.repo_path, file_path)
         if not os.path.isfile(file_path):
             return None
-        return read_file(file_path)
+        content = read_file(file_path)
+        if file_path.endswith(".html") or file_path.endswith(".htm"):
+            content = md(content, escape_underscores=False)
+        tokens = count_tokens(content)
+        MAX_TOKENS = os.environ.get('OPENAI_MAX_INPUT_TOKENS', 102400)
+        if tokens > int(MAX_TOKENS):
+            content = content[:100000]
+        return content
 
 class summarize_file_tool(agent_tool):
     """ read and summarize the file

bioguider-0.2.7/bioguider/agents/evaluation_installation_task.py

@@ -0,0 +1,160 @@
+import os
+from pathlib import Path
+import logging
+from typing import Callable, Optional
+from abc import ABC, abstractmethod
+from langchain.prompts import ChatPromptTemplate
+from langchain_openai.chat_models.base import BaseChatOpenAI
+from pydantic import BaseModel, Field
+from markdownify import markdownify as md
+
+from bioguider.agents.agent_utils import read_file
+from bioguider.utils.constants import DEFAULT_TOKEN_USAGE, ProjectMetadata
+from bioguider.rag.data_pipeline import count_tokens
+from .common_agent_2step import CommonAgentTwoSteps, CommonAgentTwoChainSteps
+from .common_agent import CommonConversation
+from ..utils.pyphen_utils import PyphenReadability
+from ..utils.gitignore_checker import GitignoreChecker
+from .evaluation_task import EvaluationTask
+from .agent_utils import read_file
+
+
+logger = logging.getLogger(__name__)
+
+EVALUATION_INSTALLATION_SYSTEM_PROMPT = """
+You are an expert in evaluating the quality of **installation instructions** in software repositories.
+Your task is to analyze the provided content of installation-related files and generate a **comprehensive, structured quality report**.
+
+---
+
+### Evaluation Criteria
+
+Please assess the installation information using the following criteria. For each, provide a concise evaluation and specific feedback:
+
+1. **Ease of Access**
+   * Is the installation information clearly presented and easy to locate within the repository?
+   * Is it included in a top-level README, a dedicated INSTALL.md file, or other accessible locations?
+
+2. **Clarity of Dependency Specification**
+   * Are all software and library dependencies clearly listed?
+   * Are installation methods (e.g., `pip`, `conda`, `apt`) for those dependencies explicitly provided?
+
+3. **Hardware Requirements**
+   * Does the documentation specify hardware needs (e.g., GPU, memory, OS) if relevant?
+
+4. **Step-by-Step Installation Guide**
+   * Is there a clear, ordered set of instructions for installing the software?
+   * Are example commands or configuration steps provided to help users follow along?
+
+---
+
+### Output Format
+
+Your response **must exactly follow** the structure below:
+
+```
+**FinalAnswer**
+**Overall Score:** [Poor / Fair / Good / Excellent]  
+**Ease of Access:** <your comments>  
+**Clarity of Dependency Specification:** <your comments>  
+**Hardware Requirements:** <your comments>  
+**Installation Guide:** <your comments>  
+```
+
+---
+
+### Installation Files Provided:
+{installation_file_contents}
+
+"""
+
+class EvaluationInstallationResult(BaseModel):
+    ease_of_access: Optional[str]=Field(description="Is the installation information easy to access")
+    score: Optional[str]=Field(description="An overall score, could be Poor, Fair, Good or Excellent")
+    clarity_of_dependency: Optional[str]=Field(description="Are all dependencies clearly listed")
+    hardware_requirements: Optional[str]=Field(description="Are all hardware requirements clearly specified")
+    installation_guide: Optional[str]=Field(description="Is there a clear, ordered set of instructions for installing the software")
+
+EvaluationInstallationResultSchema = {
+    "title": "EvaluationREADMEResult",
+    "type": "object",
+    "properties": {
+        "ease_of_access": {
+            "anyOf": [{"type": "string"}, {"type": "null"}],
+            "description": "Is the installation information easy to access",
+            "title": "Ease of Access"
+        },
+        "score": {
+            "anyOf": [{"type": "string"}, {"type": "null"}],
+            "description": "An overall score, could be Poor, Fair, Good or Excellent",
+            "title": "Score"
+        },
+        "clarity_of_dependency": {
+            "anyOf": [{"type": "string"}, {"type": "null"}],
+            "description": "Are all dependencies clearly listed",
+            "title": "Clarity of Dependency",
+        },
+        "hardware_requirements": {
+            "anyOf": [{"type": "string"}, {"type": "null"}],
+            "description": "Are all hardware requirements clearly specified",
+            "title": "Hardware Requirements"
+        },
+        "installation_guide": {
+            "anyOf": [{"type": "string"}, {"type": "null"}],
+            "description": "Is there a clear, ordered set of instructions for installing the software",
+            "title": "Installation Guide"
+        }
+    },
+    "required": ["ease_of_access", "score", "clarity_of_dependency", "hardware_requirements", "installation_guide"]
+}
+
+class EvaluationInstallationTask(EvaluationTask):
+    def __init__(
+        self, 
+        llm, 
+        repo_path,
+        gitignore_path, 
+        meta_data = None, 
+        step_callback = None,
+    ):
+        super().__init__(llm, repo_path, gitignore_path, meta_data, step_callback)
+
+    def _evaluate(self, files: list[str] | None = None):
+        if files is None or len(files) == 0:
+            return None
+        
+        files_content = ""
+        MAX_TOKENS = os.environ.get("OPENAI_MAX_INPUT_TOKENS", 102400)
+        for f in files:
+            if f.endswith(".html") or f.endswith(".htm"):
+                html = read_file(os.path.join(self.repo_path, f))
+                content = md(html, escape_underscores=False)
+            else:
+                content = read_file(os.path.join(self.repo_path, f))
+            if count_tokens(content) > int(MAX_TOKENS):
+                content = content[:100000]
+            files_content += f"""
+{f} content:
+{content}
+
+"""
+        system_prompt = ChatPromptTemplate.from_template(EVALUATION_INSTALLATION_SYSTEM_PROMPT).format(
+            installation_file_contents=files_content
+        )
+        agent = CommonAgentTwoChainSteps(llm=self.llm)
+        res, _, token_usage, reasoning_process = agent.go(
+            system_prompt=system_prompt,
+            instruction_prompt="Before arriving at the conclusion, clearly explain your reasoning step by step. Now, let's begin the evaluation.",
+            schema=EvaluationInstallationResultSchema,
+        )
+        res = EvaluationInstallationResult(**res)
+        evaluation = {
+            "score": res.score,
+            "ease_of_access": res.ease_of_access,
+            "hardware_requirements": res.hardware_requirements,
+            "clarity_of_dependency": res.clarity_of_dependency,
+            "installation_guide": res.installation_guide,
+            "reasoning_process": reasoning_process,
+        }
+        return evaluation, token_usage
+        

{bioguider-0.2.5 → bioguider-0.2.7}/bioguider/agents/evaluation_task.py

@@ -76,7 +76,7 @@ For each criterion below, provide a brief assessment followed by specific, actio
   * Project-Level README: Yes / No
   * **Score:** <number from 0 to 100>  
   * **Key Strengths**: <brief summary of the README's strongest points in 2-3 sentences> 
-  * **Improvement Suggestions:**
+  * **Overall Improvement Suggestions:**
     - "Original text snippet 1" - Improving comment 1  
     - "Original text snippet 2" - Improving comment 2  
     - ...
@@ -124,7 +124,12 @@ For each criterion below, provide a brief assessment followed by specific, actio
 **Final Answer**
   The final answer **must exactly match** the following format:
  * Project-Level README: Yes / No
- * Overall Assessment: provide a final, overall assessment of the README file's quality, summarizing the key strengths and areas for improvement.
+ * **Score:** <number from 0 to 100>  
+  * **Key Strengths**: <brief summary of the README's strongest points in 2-3 sentences> 
+  * **Overall Improvement Suggestions:**
+    - "Original text snippet 1" - Improving comment 1  
+    - "Original text snippet 2" - Improving comment 2  
+    - ...
 ---
 
 ### **README path:**
@@ -243,7 +248,9 @@ class EvaluationREADMETask(EvaluationTask):
                 readme_evaluations[readme_file] = {
                     "evaluation": {
                         "project_level": "/" in readme_file,
-                        "overall_assessment": f"{readme_file} is an empty file."
+                        "score": 0,
+                        "key_strengths": f"{readme_file} is an empty file.",
+                        "overall_improvement_suggestions": f"{readme_file} is an empty file.",
                     },
                     "reasoning_process": f"{readme_file} is an empty file.",
                 }
@@ -277,7 +284,7 @@ class EvaluationREADMETask(EvaluationTask):
                     "project_level": response.project_level,
                     "score": response.score,
                     "key_strengths": response.key_strengths,
-                    "overall_assessment": response.overall_improvement_suggestions,
+                    "overall_improvement_suggestions": response.overall_improvement_suggestions,
                 }, 
                 "reasoning_process": reasoning_process
             }

{bioguider-0.2.5 → bioguider-0.2.7}/bioguider/agents/prompt_utils.py

@@ -165,7 +165,7 @@ If **any one** of these is present, the document should be classified as Install
         "important_instructions": """- Give priority to analyzing README file that contain installation instructions and the files whose names include **"install"** or **"setup"**.
 - If multiple files are found, select the most comprehensive one that covers the installation process.
 - The total number of collected files should **not exceed 3**.
-- Make sure to include **only one installation instruction file**, selecting the most comprehensive and representative one.
+- Identify and select **no more than three** installation instruction files — choose the most comprehensive and representative ones.
 """
     },
     "License": {

{bioguider-0.2.5 → bioguider-0.2.7}/bioguider/managers/evaluation_manager.py

@@ -10,6 +10,7 @@ from ..rag.rag import RAG
 from ..utils.file_utils import parse_repo_url
 from ..database.summarized_file_db import SummarizedFilesDb
 from ..agents.evaluation_task import EvaluationREADMETask
+from ..agents.evaluation_installation_task import EvaluationInstallationTask
 from ..agents.collection_task import CollectionTask
 
 class EvaluationManager:
@@ -84,6 +85,30 @@ class EvaluationManager:
         if s is None or 'final_answer' not in s:
             return None
         
+    def evaluate_installation(self):
+        task = CollectionTask(
+            llm=self.llm,
+            step_callback=self.step_callback,
+        )
+        task.compile(
+            repo_path=self.rag.repo_dir,
+            gitignore_path=Path(self.rag.repo_dir, ".gitignore"),
+            db=self.summary_file_db,
+            goal_item=CollectionGoalItemEnum.Installation.name,
+        )
+        files = task.collect()
+        if files is None or len(files) == 0:
+            return None
+        evaluation_task = EvaluationInstallationTask(
+            llm=self.llm,
+            repo_path=self.rag.repo_dir,
+            gitignore_path=Path(self.rag.repo_dir, ".gitignore"),
+            meta_data=self.project_metadata,
+            step_callback=self.step_callback,
+        )
+        evaluation = evaluation_task.evaluate(files)
+        return evaluation, files
+        
     def _find_readme_files(self) -> list[str]:
         """
         Search for a README file in the repository directory.

{bioguider-0.2.5 → bioguider-0.2.7}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "bioguider"
-version = "0.2.5"
+version = "0.2.7"
 description = "An AI-Powered package to help biomedical developers to generate clear documentation"
 authors = [
     "Cankun Wang <Cankun.Wang@osumc.edu>",
@@ -45,6 +45,7 @@ faiss-cpu = "^1.11.0"
 binaryornot = "^0.4.4"
 textstat = "^0.7.6"
 pyphen = "^0.17.2"
+markdownify = "^1.1.0"
 
 
 [tool.poetry.group.dev.dependencies]