bioguider 0.2.19__py3-none-any.whl → 0.2.20__py3-none-any.whl

bioguider/managers/generation_test_manager.py

@@ -0,0 +1,74 @@
+from __future__ import annotations
+
+import os
+import json
+from typing import Tuple
+
+from bioguider.generation.llm_injector import LLMErrorInjector
+from bioguider.generation.test_metrics import evaluate_fixes
+from bioguider.managers.generation_manager import DocumentationGenerationManager
+from bioguider.agents.agent_utils import read_file, write_file
+
+
+class GenerationTestManager:
+    def __init__(self, llm, step_callback):
+        self.llm = llm
+        self.step_output = step_callback
+
+    def print_step(self, name: str, out: str | None = None):
+        if self.step_output:
+            self.step_output(step_name=name, step_output=out)
+
+    def run_quant_test(self, report_path: str, baseline_repo_path: str, tmp_repo_path: str) -> str:
+        self.print_step("QuantTest:LoadBaseline", baseline_repo_path)
+        baseline_readme_path = os.path.join(baseline_repo_path, "README.md")
+        baseline = read_file(baseline_readme_path) or ""
+
+        self.print_step("QuantTest:Inject")
+        injector = LLMErrorInjector(self.llm)
+        corrupted, inj_manifest = injector.inject(baseline, min_per_category=3)
+
+        # write corrupted into tmp repo path
+        os.makedirs(tmp_repo_path, exist_ok=True)
+        corrupted_readme_path = os.path.join(tmp_repo_path, "README.md")
+        write_file(corrupted_readme_path, corrupted)
+        inj_path = os.path.join(tmp_repo_path, "INJECTION_MANIFEST.json")
+        with open(inj_path, "w", encoding="utf-8") as fobj:
+            json.dump(inj_manifest, fobj, indent=2)
+
+        self.print_step("QuantTest:Generate")
+        gen = DocumentationGenerationManager(self.llm, self.step_output)
+        out_dir = gen.run(report_path=report_path, repo_path=tmp_repo_path)
+
+        # read revised
+        revised_readme_path = os.path.join(out_dir, "README.md")
+        revised = read_file(revised_readme_path) or ""
+
+        self.print_step("QuantTest:Evaluate")
+        results = evaluate_fixes(baseline, corrupted, revised, inj_manifest)
+        # write results
+        with open(os.path.join(out_dir, "GEN_TEST_RESULTS.json"), "w", encoding="utf-8") as fobj:
+            json.dump(results, fobj, indent=2)
+        # simple md report
+        lines = ["# Quantifiable Generation Test Report\n"]
+        lines.append("## Metrics by Category\n")
+        for cat, m in results["per_category"].items():
+            lines.append(f"- {cat}: {m}")
+        lines.append("\n## Notes\n")
+        lines.append("- Three versions saved in this directory: README.original.md, README.corrupted.md, README.md (fixed).")
+        with open(os.path.join(out_dir, "GEN_TEST_REPORT.md"), "w", encoding="utf-8") as fobj:
+            fobj.write("\n".join(lines))
+        # Save versioned files into output dir
+        write_file(os.path.join(out_dir, "README.original.md"), baseline)
+        write_file(os.path.join(out_dir, "README.corrupted.md"), corrupted)
+        # Copy injection manifest
+        try:
+            with open(inj_path, "r", encoding="utf-8") as fin:
+                with open(os.path.join(out_dir, "INJECTION_MANIFEST.json"), "w", encoding="utf-8") as fout:
+                    fout.write(fin.read())
+        except Exception:
+            pass
+        self.print_step("QuantTest:Done", out_dir)
+        return out_dir
+
+

bioguider/utils/code_structure_builder.py

@@ -0,0 +1,42 @@
+from pathlib import Path
+import logging
+
+from .gitignore_checker import GitignoreChecker
+from .file_handler import FileHandler
+from ..database.code_structure_db import CodeStructureDb
+
+logger = logging.getLogger(__name__)
+
+class CodeStructureBuilder:
+    def __init__(
+        self, 
+        repo_path: str, 
+        gitignore_path: str,
+        code_structure_db: CodeStructureDb,
+    ):
+        self.repo_path = repo_path
+        self.gitignore_checker = GitignoreChecker(repo_path, gitignore_path)
+        self.file_handler = FileHandler(repo_path)
+        self.code_structure_db = code_structure_db
+
+    def build_code_structure(self):
+        files = self.gitignore_checker.check_files_and_folders()
+        for file in files:
+            if not file.endswith(".py"):
+                continue
+            logger.info(f"Building code structure for {file}")
+            file_handler = FileHandler(Path(self.repo_path) / file)
+            functions_and_classes = file_handler.get_functions_and_classes()
+            # fixme: currently, we don't extract reference graph for each function or class
+            for function_or_class in functions_and_classes:
+                self.code_structure_db.insert_code_structure(
+                    function_or_class[0], # name
+                    file,
+                    function_or_class[2], # start line number
+                    function_or_class[3], # end line number
+                    function_or_class[1], # parent name
+                    function_or_class[4], # doc string
+                    function_or_class[5], # params
+                )
+        
+

bioguider/utils/file_handler.py

@@ -0,0 +1,65 @@
+import ast
+import os
+
+class FileHandler:
+    def __init__(self, file_path: str):
+        self.file_path = file_path
+
+    def get_functions_and_classes(self) -> list[str]:
+        """
+        Get the functions and classes in a given file.
+        Returns a list of tuples, each containing:
+        1. the function or class name,
+        2. parent name,
+        3. start line number,
+        4. end line number,
+        5. doc string,
+        6. params.
+        """
+        with open(self.file_path, 'r') as f:
+            tree = ast.parse(f.read())
+            functions_and_classes = []
+            for node in tree.body:
+                if isinstance(node, ast.FunctionDef) or isinstance(node, ast.ClassDef):
+                    start_lineno = node.lineno
+                    end_lineno = self.get_end_lineno(node)
+                    doc_string = ast.get_docstring(node)
+                    params = (
+                        [arg.arg for arg in node.args.args] if "args" in dir(node) else []
+                    )
+                    parent = None
+                    functions_and_classes.append((node.name, parent, start_lineno, end_lineno, doc_string, params))
+                    for child in node.body:
+                        if isinstance(child, ast.FunctionDef):
+                            start_lineno = child.lineno
+                            end_lineno = self.get_end_lineno(child)
+                            doc_string = ast.get_docstring(child)
+                            params = (
+                                [arg.arg for arg in child.args.args] if "args" in dir(child) else []
+                            )
+                            parent = node.name
+                            functions_and_classes.append((child.name, parent, start_lineno, end_lineno, doc_string, params))
+            return functions_and_classes
+    
+    def get_imports(self) -> list[str]:
+        pass
+
+    def get_end_lineno(self, node):
+        """
+        Get the end line number of a given node.
+
+        Args:
+            node: The node for which to find the end line number.
+
+        Returns:
+            int: The end line number of the node. Returns -1 if the node does not have a line number.
+        """
+        if not hasattr(node, "lineno"):
+            return -1  # 返回-1表示此节点没有行号
+
+        end_lineno = node.lineno
+        for child in ast.iter_child_nodes(node):
+            child_end = getattr(child, "end_lineno", None) or self.get_end_lineno(child)
+            if child_end > -1:  # 只更新当子节点有有效行号时
+                end_lineno = max(end_lineno, child_end)
+        return end_lineno

{bioguider-0.2.19.dist-info → bioguider-0.2.20.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: bioguider
-Version: 0.2.19
+Version: 0.2.20
 Summary: An AI-Powered package to help biomedical developers to generate clear documentation
 License: MIT
 Author: Cankun Wang

{bioguider-0.2.19.dist-info → bioguider-0.2.20.dist-info}/RECORD

@@ -2,16 +2,21 @@ bioguider/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bioguider/agents/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bioguider/agents/agent_task.py,sha256=TL0Zx8zOmiAVslmNbfMPQ38qTQ73QospY6Dwrwf8POg,2890
 bioguider/agents/agent_tools.py,sha256=r21wHV6a-Ic2T0dk4YzA-_d7PodHPM3GzRxJqv-llSw,7286
-bioguider/agents/agent_utils.py,sha256=Om1xwAmRdHFC4jkbtaBM3p309z5KFjMX1a6BFVPXX68,14449
-bioguider/agents/collection_execute_step.py,sha256=Ev4BLjjmBdsc52M1zrq7QK8g7fsffDkSxu-jN2rvedw,5614
+bioguider/agents/agent_utils.py,sha256=SoYFc5oZGY2QqTDHR1DCgBPiwQpigUOANHmlUCUPzu4,14683
+bioguider/agents/collection_execute_step.py,sha256=jE_oSQZI5WDaz0bJjUWoAfqWfVbGUqN--cvITSWCGiI,5614
 bioguider/agents/collection_observe_step.py,sha256=n863HrbVANQVeltffjS2zXv-AfVErC8ZEMfb_78hafk,5140
 bioguider/agents/collection_plan_step.py,sha256=Nn0f8AOkEDCDtnhaqE7yCQoi7PVpsHmiUcsIqC0T0dQ,5956
 bioguider/agents/collection_task.py,sha256=MjpTYiQQYUpmQf2UOn-dOCZU3kxypc4uOnzd15wb1Ow,7882
 bioguider/agents/collection_task_utils.py,sha256=_e2EebYhl-UYjZ0rHNf2-p32YlstBSffv32suiuT9LI,5386
 bioguider/agents/common_agent.py,sha256=TpfxbYskwuwWrjs1g9RaG7sdA5rOLdiVac7If7uK2sg,4558
 bioguider/agents/common_agent_2step.py,sha256=rGiDzUkmmUIFnmJJxzXK5M5BfIyINHXLZ0pmPRUVqQg,7911
-bioguider/agents/common_conversation.py,sha256=VrnQoPXXK3xeFjeP4-HH-yD9qmw-o9v93LFtNbYey6E,1655
+bioguider/agents/common_conversation.py,sha256=_9l6SunRmOZ_3R4Q8CTO9oE_qmP7aYYKFX1EiFBIrm8,2589
 bioguider/agents/common_step.py,sha256=GdOCbmj1pwh4etg-futVFYVDQuoUG89DnIrw-B6QbzM,2594
+bioguider/agents/consistency_collection_execute_step.py,sha256=SBZjdaIc4AX6ljedd8TsJC-1GoYeI16jaqIcfDnTezs,4938
+bioguider/agents/consistency_collection_observe_step.py,sha256=XghjlSMBpq1Va68jRP8_-ZG0O7UlZy1Qhwai9c0LNoo,5029
+bioguider/agents/consistency_collection_plan_step.py,sha256=NUl0Nl39O5EFibJQQB1oMIVY_Tikh75YNVbyt7ryw6s,5389
+bioguider/agents/consistency_collection_task.py,sha256=WP3GL4iPJ38b5F1zZMHmJALx0pRDcF-9JAmz3OHqi6o,5261
+bioguider/agents/consistency_collection_task_utils.py,sha256=BnpBXmpxmmr8NRaBw5sACipP0mMtj1Qpe2AvcMlX_cg,6143
 bioguider/agents/dockergeneration_execute_step.py,sha256=F92jDlkc6KjAvTkX7q1FsCYP8J15SCaNgmwh3YPqfDo,6500
 bioguider/agents/dockergeneration_observe_step.py,sha256=Bo5Td0fzMYLbLki0FvwamzqRFOy4eu3AvIUa8oFApE4,6131
 bioguider/agents/dockergeneration_plan_step.py,sha256=SB8tQM9PkIKsD2o1DFD7bedcxz6r6hSy8n_EVK60Fz0,7235
@@ -20,32 +25,52 @@ bioguider/agents/dockergeneration_task_utils.py,sha256=v7emqrJlVW-A5ZdLmPSdiaMSK
 bioguider/agents/evaluation_installation_task.py,sha256=0RNH5NV7YKwn_we_d3IjnFf_ee9IPCEQ_olebq2y8Ng,12130
 bioguider/agents/evaluation_readme_task.py,sha256=pi3oAGJgZhJgJG1xLgiobrk3Uy2a_JIarD5QSPBkmHA,30647
 bioguider/agents/evaluation_submission_requirements_task.py,sha256=J_6C-M2AfYue2C-gWBHl7KqGrTBuFBn9zmMV5vSRk-U,7834
-bioguider/agents/evaluation_task.py,sha256=4UrcUCy8UIVLd1NpBpqGgmudqeSK3wWI3Jm8LEYySbY,12661
+bioguider/agents/evaluation_task.py,sha256=c3kvc3xgzGbT0C2KpkE-zHBvaxx9zKdmpLzeKHddrds,12690
+bioguider/agents/evaluation_userguide_prompts.py,sha256=eyJUx5nUr8v9k0B5GpKDaX2dBxSLVZGA0fwOWS4Uiow,7154
+bioguider/agents/evaluation_userguide_task.py,sha256=AJ7y5WSMs4EqXpQzyAu_5NPmWK0jl3AeHhRXxdZgpz0,8266
 bioguider/agents/identification_execute_step.py,sha256=w3IjL8f2WiHCyiLjVSoySnIAXpi1-hK1DLKCnXbAN2Y,5587
 bioguider/agents/identification_observe_step.py,sha256=Me5mhEM4e7FGnVFcluNtqfhIxzng6guGIu39xi1TrS8,4341
 bioguider/agents/identification_plan_step.py,sha256=owsTK1NZIuiZL7QPVknJyp9TBRK-mhnuf2RwK4YzaxU,5442
 bioguider/agents/identification_task.py,sha256=bTbovxxQVpO1TcdcQAxDxwPISuAcXndO7zsvHpJSb64,10147
 bioguider/agents/identification_task_utils.py,sha256=Lf0Rj0L0KSiyJmPAgeSz0vLUFQr6TSFuzgufimEN4H0,630
 bioguider/agents/peo_common_step.py,sha256=iw2c1h7X11WJzSE2tSRg0UAoXH0QOlQDxW9CCzSVMOY,2677
-bioguider/agents/prompt_utils.py,sha256=EuR7NkB7PcCgNroeX91fHUSK2X6mLS6vK7RnzyQHAHI,17593
+bioguider/agents/prompt_utils.py,sha256=kTeK0LOsxpEIpQ5PMXY0DOHXJ5Z6ryu0UuJIO4ga9Rg,17743
 bioguider/agents/python_ast_repl_tool.py,sha256=o7-4P1h8jS8ikhGSA4CI_OWQ2a0Eg5tEdmuAp_qrO-0,2519
 bioguider/agents/rag_collection_task.py,sha256=r_jPAMjQcC7dIydKxX77UuMqjJ3MiVKswNZ-yNw7yx8,5199
 bioguider/conversation.py,sha256=DIvk_d7pz_guuORByK1eaaF09FAK-8shcNTrbSUHz9Y,1779
+bioguider/database/code_structure_db.py,sha256=drYkoQa7mDexCpulG3885Gwcnu92lnHNRlWT8byjF-A,16630
 bioguider/database/summarized_file_db.py,sha256=gGubM5vH3SN_1vouKO0LemvE_EfyjinAScRhlvxJsi0,4654
-bioguider/managers/evaluation_manager.py,sha256=O8mxrAGllDIkcQVsrRrqxH0eyJHwtoSauWrPe_F7qqU,4778
+bioguider/generation/__init__.py,sha256=esV02QgCsY67-HBwSHDbA5AcbKzNRIT3wDwwh6N4OFM,945
+bioguider/generation/change_planner.py,sha256=1u_1tcVwggVw6QebXvPJVULYnRAIkKMteFxbnf42aAw,6964
+bioguider/generation/document_renderer.py,sha256=67-XGFGNe8WgHJpqkHLMecgk2OeV8a5cJRAeDTLsXG4,1806
+bioguider/generation/llm_cleaner.py,sha256=aMxc3j6IH_9mCfBjxY1bn0bAUtUkUoHjumobzGkN0h8,1422
+bioguider/generation/llm_content_generator.py,sha256=45z_SWaoL0Z4hBFhYJHT-_L7mFwaLgvGocQiXIBmCpw,3332
+bioguider/generation/llm_injector.py,sha256=Q6F9IshTlZ7KUiRvOl1gT4RKQn-AtD1QiN7aK0gba4A,13083
+bioguider/generation/models.py,sha256=FEfXvszF2sGOYa98Jn4VDp6TyktJL4cD6LQFhx9rmsk,2328
+bioguider/generation/output_manager.py,sha256=V0eXYG9iaWxUjU44eLLX5mQcfjOz5tAKa_X0rkTc5Aw,1935
+bioguider/generation/repo_reader.py,sha256=ivTURU61fR8er4ev7gSpOxER3FJv2d9GAx_X5JoVTvQ,1177
+bioguider/generation/report_loader.py,sha256=KtJ6JHGPDL-PQoucU8hkVzMSz-B0bHbF2WShjuz2fY8,5102
+bioguider/generation/style_analyzer.py,sha256=Vn9FAK1qJBNLolLC1tz362k4UBaPl107BlvkQc8pV2I,983
+bioguider/generation/suggestion_extractor.py,sha256=tfkyWtdbAo-maLCF_wqwBXyh93yjulvDY17FuvTnTjk,7611
+bioguider/generation/test_metrics.py,sha256=fG6H1jVikHEx1YvN5Ds4QbVinudJ5OEYkzrV760oLLQ,3766
+bioguider/managers/evaluation_manager.py,sha256=obsesQdEgxBzTd1Re00kanH1d5_4_vxzXI_ndICgedg,5834
+bioguider/managers/generation_manager.py,sha256=EbAJSvUz-SIriVlozuJ6wa5_1aIbbFfpgg3c9Vcz34g,7615
+bioguider/managers/generation_test_manager.py,sha256=0ty8IibdfN90Oj6M6lkYbxASnQxHYb9I3w9eG7hvEsQ,3270
 bioguider/rag/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bioguider/rag/config.py,sha256=5g4IqTzgyfZfax9Af9CTkXShgItPOt4_9TEMSekCPik,4602
 bioguider/rag/data_pipeline.py,sha256=bkJ2IUCgPx_OL2uZtPd6cIBor2VFZEIfGd5kVlmiPjw,27292
 bioguider/rag/embedder.py,sha256=jofR8hOj3Aj2IyBQ9y6FeAc84tgq5agbIfCGyFxYpJ8,650
 bioguider/rag/rag.py,sha256=JFPwrJlKDSyd3U3Gce_NSxI5343eNUbqPG9Fs5Pfoq0,4696
 bioguider/settings.py,sha256=BD_iz9aYarxmWUl0XaKl4-D4oTXMhFzljsXLNn2phis,3143
+bioguider/utils/code_structure_builder.py,sha256=TBrsmlUPEwqpi5bSVLrOEwFK1dQ49goYJ1EGYmRrtZ8,1607
 bioguider/utils/constants.py,sha256=Yzo9oFG4P4C83vKl22RsWkIDa4dcitvcCIceusNPIRQ,8928
 bioguider/utils/default.gitignore,sha256=XjPdyO2KV8z8iyuqluaNR_70tBQftMpyKL8HboVNyeI,1605
+bioguider/utils/file_handler.py,sha256=xTrJBQ7VY5baIJPpX7YJP0oniPO3y9ltzayqUOQXJ5g,2644
 bioguider/utils/file_utils.py,sha256=9VfAHsz1UkFPtzAmvWZvPl1TMaKIYNjNlLgsfB8tNjg,3683
 bioguider/utils/gitignore_checker.py,sha256=pOYUwsS9D5014LxcZb0cj3s2CAYaD2uF_pYJpaNKcho,6532
 bioguider/utils/pyphen_utils.py,sha256=cdZc3qphkvMDeL5NiZ8Xou13M_uVNP7ifJ-FwxO-0BE,2680
 bioguider/utils/utils.py,sha256=aWtgdvB04gEiJTfQNK4aQPO1mxv2zRZTbDaGUBy9DFc,2275
-bioguider-0.2.19.dist-info/LICENSE,sha256=qzkvZcKwwA5DuSuhXMOm2LcO6BdEr4V7jwFZVL2-jL4,1065
-bioguider-0.2.19.dist-info/METADATA,sha256=YvAwNzOyncowyrbKcoKBSEeIVLPk8UiRK-Y4vihrUUY,1868
-bioguider-0.2.19.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-bioguider-0.2.19.dist-info/RECORD,,
+bioguider-0.2.20.dist-info/LICENSE,sha256=qzkvZcKwwA5DuSuhXMOm2LcO6BdEr4V7jwFZVL2-jL4,1065
+bioguider-0.2.20.dist-info/METADATA,sha256=jAGUMMHuLYW5_lyvc0tYcvKGYlB9YUG1hqT3Kjaqii0,1868
+bioguider-0.2.20.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+bioguider-0.2.20.dist-info/RECORD,,