logdetective 0.2.8__py3-none-any.whl → 0.2.10__py3-none-any.whl

logdetective/constants.py

@@ -32,9 +32,7 @@ Answer:
 """
 
 SNIPPET_PROMPT_TEMPLATE = """
-Analyse following RPM build log snippet.
-Analysis of the snippets must be in a format of [X] : [Y], where [X] is a log snippet, and [Y] is the explanation.
-Snippets themselves must not be altered in any way whatsoever.
+Analyse following RPM build log snippet. Decribe contents accurately, without speculation or suggestions for resolution.
 
 Snippet:
 
@@ -43,3 +41,22 @@ Snippet:
 Analysis:
 
 """
+
+PROMPT_TEMPLATE_STAGED = """
+Given following log snippets, their explanation, and nothing else, explain what failure, if any, occured during build of this package.
+
+Snippets are in a format of [X] : [Y], where [X] is a log snippet, and [Y] is the explanation.
+
+Snippets are delimited with '================'.
+
+Drawing on information from all snippets, provide complete explanation of the issue and recommend solution.
+
+Snippets:
+
+{}
+
+Analysis:
+
+"""
+
+SNIPPET_DELIMITER = '================'

logdetective/extractors.py

@@ -16,7 +16,7 @@ class LLMExtractor:
     A class that extracts relevant information from logs using a language model.
     """
     def __init__(self, model: Llama, n_lines: int = 2):
-        self.model =  model
+        self.model = model
         self.n_lines = n_lines
         self.grammar = LlamaGrammar.from_string(
             "root ::= (\"Yes\" | \"No\")", verbose=False)

logdetective/logdetective.py

@@ -9,6 +9,7 @@ from logdetective.extractors import LLMExtractor, DrainExtractor
 
 LOG = logging.getLogger("logdetective")
 
+
 def setup_args():
     """ Setup argument parser and return arguments. """
     parser = argparse.ArgumentParser("logdetective")

logdetective/server.py

@@ -2,7 +2,7 @@ import asyncio
 import json
 import logging
 import os
-from typing import List, Annotated
+from typing import List, Annotated, Dict
 
 from llama_cpp import CreateCompletionResponse
 from fastapi import FastAPI, HTTPException, Depends, Header
@@ -10,10 +10,13 @@ from fastapi.responses import StreamingResponse
 from pydantic import BaseModel
 import requests
 
-from logdetective.constants import PROMPT_TEMPLATE, SNIPPET_PROMPT_TEMPLATE
+from logdetective.constants import (
+    PROMPT_TEMPLATE, SNIPPET_PROMPT_TEMPLATE,
+    PROMPT_TEMPLATE_STAGED, SNIPPET_DELIMITER)
 from logdetective.extractors import DrainExtractor
 from logdetective.utils import validate_url, compute_certainty
 
+
 class BuildLog(BaseModel):
     """Model of data submitted to API.
     """
@@ -38,14 +41,13 @@ class StagedResponse(Response):
     explanation: CreateCompletionResponse
         https://llama-cpp-python.readthedocs.io/en/latest/api-reference/#llama_cpp.llama_types.CreateCompletionResponse
     response_certainty: float
-    snippets: list of CreateCompletionResponse
+    snippets:
+        list of dictionaries { 'snippet' : '<original_text>, 'comment': CreateCompletionResponse }
     """
-    snippets: List[CreateCompletionResponse]
-
+    snippets: List[Dict[str, str | CreateCompletionResponse]]
 
 LOG = logging.getLogger("logdetective")
 
-
 LLM_CPP_HOST = os.environ.get("LLAMA_CPP_HOST", "localhost")
 LLM_CPP_SERVER_ADDRESS = f"http://{LLM_CPP_HOST}"
 LLM_CPP_SERVER_PORT = os.environ.get("LLAMA_CPP_SERVER_PORT", 8000)
@@ -53,6 +55,7 @@ LLM_CPP_SERVER_TIMEOUT = os.environ.get("LLAMA_CPP_SERVER_TIMEOUT", 600)
 LOG_SOURCE_REQUEST_TIMEOUT = os.environ.get("LOG_SOURCE_REQUEST_TIMEOUT", 60)
 API_TOKEN = os.environ.get("LOGDETECTIVE_TOKEN", None)
 
+
 def requires_token_when_set(authentication: Annotated[str | None, Header()] = None):
     """
     FastAPI Depend function that expects a header named Authentication
@@ -80,6 +83,7 @@ def requires_token_when_set(authentication: Annotated[str | None, Header()] = No
              API_TOKEN, token)
     raise HTTPException(status_code=401, detail=f"Token {token} not valid.")
 
+
 app = FastAPI(dependencies=[Depends(requires_token_when_set)])
 
 
@@ -97,7 +101,7 @@ def process_url(url: str) -> str:
         if not log_request.ok:
             raise HTTPException(status_code=400,
                                 detail="Something went wrong while getting the logs: "
-                                    f"[{log_request.status_code}] {log_request.text}")
+                                f"[{log_request.status_code}] {log_request.text}")
     else:
         LOG.error("Invalid URL received ")
         raise HTTPException(status_code=400,
@@ -118,9 +122,9 @@ def mine_logs(log: str) -> List[str]:
     LOG.debug("Log summary: \n %s", log_summary)
     LOG.info("Compression ratio: %s", ratio)
 
-
     return log_summary
 
+
 async def submit_text(text: str, max_tokens: int = 0, log_probs: int = 1, stream: bool = False,
                       model: str = "default-model"):
     """Submit prompt to LLM.
@@ -129,17 +133,17 @@ async def submit_text(text: str, max_tokens: int = 0, log_probs: int = 1, stream
     """
     LOG.info("Analyzing the text")
     data = {
-            "prompt": text,
-            "max_tokens": str(max_tokens),
-            "logprobs": str(log_probs),
-            "stream": stream,
-            "model": model}
+        "prompt": text,
+        "max_tokens": str(max_tokens),
+        "logprobs": str(log_probs),
+        "stream": stream,
+        "model": model}
 
     try:
         # Expects llama-cpp server to run on LLM_CPP_SERVER_ADDRESS:LLM_CPP_SERVER_PORT
         response = requests.post(
             f"{LLM_CPP_SERVER_ADDRESS}:{LLM_CPP_SERVER_PORT}/v1/completions",
-            headers={"Content-Type":"application/json"},
+            headers={"Content-Type": "application/json"},
             data=json.dumps(data),
             timeout=int(LLM_CPP_SERVER_TIMEOUT),
             stream=stream)
@@ -152,7 +156,7 @@ async def submit_text(text: str, max_tokens: int = 0, log_probs: int = 1, stream
             raise HTTPException(
                 status_code=400,
                 detail="Something went wrong while getting a response from the llama server: "
-                    f"[{response.status_code}] {response.text}")
+                       f"[{response.status_code}] {response.text}")
         try:
             response = json.loads(response.text)
         except UnicodeDecodeError as ex:
@@ -208,10 +212,18 @@ async def analyze_log_staged(build_log: BuildLog):
     analyzed_snippets = await asyncio.gather(
         *[submit_text(SNIPPET_PROMPT_TEMPLATE.format(s)) for s in log_summary])
 
-    final_analysis = await submit_text(
-        PROMPT_TEMPLATE.format([e["choices"][0]["text"] for e in analyzed_snippets]))
+    analyzed_snippets = [
+        {"snippet":e[0], "comment":e[1]} for e in zip(log_summary, analyzed_snippets)]
+
+    final_prompt = PROMPT_TEMPLATE_STAGED.format(
+        f"\n{SNIPPET_DELIMITER}\n".join([
+            f"[{e["snippet"]}] : [{e["comment"]["choices"][0]["text"]}]"
+        for e in analyzed_snippets]))
+
+    final_analysis = await submit_text(final_prompt)
 
     certainty = 0
+
     if "logprobs" in final_analysis["choices"][0]:
         try:
             certainty = compute_certainty(

logdetective/utils.py

@@ -80,6 +80,8 @@ def compute_certainty(probs: List[Dict[str, float] | None]) -> float:
     In this case it's just a matter of applying inverse operation exp.
     Of course that leaves you with a value in range <0, 1> so it needs to be multiplied by 100.
     Simply put, this is the most straightforward way to get the numbers out.
+
+    This function is used in the server codebase.
     """
 
     top_logprobs = [

{logdetective-0.2.8.dist-info → logdetective-0.2.10.dist-info}/METADATA

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.3
 Name: logdetective
-Version: 0.2.8
+Version: 0.2.10
 Summary: Log using LLM AI to search for build/test failures and provide ideas for fixing these.
 License: Apache-2.0
 Author: Jiri Podivin
@@ -22,7 +22,7 @@ Provides-Extra: server
 Requires-Dist: drain3 (>=0.9.11,<0.10.0)
 Requires-Dist: huggingface-hub (>0.23.2)
 Requires-Dist: llama-cpp-python (>0.2.56,!=0.2.86)
-Requires-Dist: numpy (>=1.26.0,<2.0.0)
+Requires-Dist: numpy (>=1.26.0)
 Requires-Dist: requests (>0.2.31)
 Project-URL: homepage, https://github.com/fedora-copr/logdetective
 Project-URL: issues, https://github.com/fedora-copr/logdetective/issues

logdetective-0.2.10.dist-info/RECORD

@@ -0,0 +1,12 @@
+logdetective/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+logdetective/constants.py,sha256=6XekuU7sbkY1Pmu4NJajgFbJ0no8PQ3DxQm8NeLKtjE,1383
+logdetective/drain3.ini,sha256=ni91eCT1TwTznZwcqWoOVMQcGEnWhEDNCoTPF7cfGfY,1360
+logdetective/extractors.py,sha256=xfan_dbGCrLH4cguJ2F6W6UkxXMz24Vob39r5-GsNV8,3102
+logdetective/logdetective.py,sha256=j34E_udG7dDY5usWmsJ0q767-Fay3K5Aw_lBZUORPSI,4373
+logdetective/server.py,sha256=ryk6bQ4xV4CTlenxiI9SgOOc0bE_GD_z46Ooo7PjFCs,9596
+logdetective/utils.py,sha256=Q51mjbeevD6fIF84QYuc1kgC768a38zCJkJwn7TuJA0,4845
+logdetective-0.2.10.dist-info/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+logdetective-0.2.10.dist-info/METADATA,sha256=8qmL187ApXgT57eXP_r2-4F732EyH5YTFiGYbcmVIQw,9791
+logdetective-0.2.10.dist-info/WHEEL,sha256=IYZQI976HJqqOpQU6PHkJ8fb3tMNBFjg-Cn-pwAbaFM,88
+logdetective-0.2.10.dist-info/entry_points.txt,sha256=3K_vXja6PmcA8sNdUi63WdImeiNhVZcEGPTaoJmltfA,63
+logdetective-0.2.10.dist-info/RECORD,,

{logdetective-0.2.8.dist-info → logdetective-0.2.10.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: poetry-core 1.9.1
+Generator: poetry-core 2.0.1
 Root-Is-Purelib: true
 Tag: py3-none-any

logdetective-0.2.8.dist-info/RECORD

@@ -1,12 +0,0 @@
-logdetective/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-logdetective/constants.py,sha256=1Ls2VJXb7NwSgi_HmTOA1c52K16SZIeDYBXlvBJ07zU,991
-logdetective/drain3.ini,sha256=ni91eCT1TwTznZwcqWoOVMQcGEnWhEDNCoTPF7cfGfY,1360
-logdetective/extractors.py,sha256=eRizRiKhC3MPTHXS5nlRKcEudEaqct7G28V1bZYGkqI,3103
-logdetective/logdetective.py,sha256=f7ASCJg_Yt6VBFieXBYgQYdenfXjC60ZdLHhzQHideI,4372
-logdetective/server.py,sha256=UDH7LEFZ-rnIrrnZDZTcYluom1XnsvYG1b5Fc2xiivs,9226
-logdetective/utils.py,sha256=nTbaDVEfbHVQPTZe58T04HHZ6JWUJ1PonRRnzGX8hY0,4794
-logdetective-0.2.8.dist-info/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
-logdetective-0.2.8.dist-info/METADATA,sha256=gFcfcgLCOQXlQu4eWbHzGzaP8z7-yMwQ36-jFJsSg7c,9797
-logdetective-0.2.8.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-logdetective-0.2.8.dist-info/entry_points.txt,sha256=3K_vXja6PmcA8sNdUi63WdImeiNhVZcEGPTaoJmltfA,63
-logdetective-0.2.8.dist-info/RECORD,,