logdetective 0.10.0__py3-none-any.whl → 0.11.2__py3-none-any.whl

logdetective/server/database/models/merge_request_jobs.py

@@ -1,6 +1,6 @@
 import enum
 import datetime
-from typing import Optional, List
+from typing import Optional, List, Tuple
 
 import backoff
 
@@ -585,3 +585,19 @@ class Reactions(Base):
             with transaction(commit=True) as session:
                 session.delete(reaction)
                 session.flush()
+
+    @classmethod
+    def get_since(
+        cls, time: datetime.datetime
+    ) -> List[Tuple[datetime.datetime, "Comments"]]:
+        """Get all the reactions on comments created after the given time
+        and the comment creation time."""
+        with transaction(commit=False) as session:
+            reactions = (
+                session.query(Comments.created_at, cls)
+                .join(Comments, cls.comment_id == Comments.id)
+                .filter(Comments.created_at > time)
+                .all()
+            )
+
+            return reactions

logdetective/server/llm.py

@@ -1,6 +1,7 @@
 import os
 import asyncio
 import json
+import random
 from typing import List, Tuple, Dict, Any, Union
 
 import backoff
@@ -102,7 +103,7 @@ def should_we_giveup(exc: aiohttp.ClientResponseError) -> bool:
     > a truthy value if the exception should not be retried
     """
     LOG.info("Should we give up on retrying error %s", exc)
-    return exc.status < 500
+    return exc.status < 400
 
 
 def we_give_up(details: backoff._typing.Details):
@@ -110,14 +111,16 @@ def we_give_up(details: backoff._typing.Details):
     retries didn't work (or we got a different exc)
     we give up and raise proper 500 for our API endpoint
     """
+    LOG.error("Last exception: %s", details["exception"])
     LOG.error("Inference error: %s", details["args"])
     raise HTTPException(500, "Request to the inference API failed")
 
 
 @backoff.on_exception(
-    backoff.expo,
+    lambda: backoff.constant([10, 30, 120]),
     aiohttp.ClientResponseError,
-    max_tries=3,
+    max_tries=4,  # 4 tries and 3 retries
+    jitter=lambda wait_gen_value: random.uniform(wait_gen_value, wait_gen_value + 30),
     giveup=should_we_giveup,
     raise_on_giveup=False,
     on_giveup=we_give_up,

logdetective/server/metric.py

@@ -84,7 +84,24 @@ def update_metrics(
 
 def track_request(name=None):
     """
-    Decorator to track requests metrics
+    Decorator to track requests/responses metrics
+
+    On entering the decorated function, it registers the time for the request
+    and saves the passed log content.
+    On exiting the decorated function, it registers the time for the response
+    and saves the generated response.
+
+    Use it to decorate server endpoints that generate a llm response
+    as in the following example:
+
+    >>> @app.post("/analyze", response_model=Response)
+    >>> @track_request()
+    >>> async def analyze_log(build_log)
+    >>>     pass
+
+    Warning: the decorators' order is important!
+    The function returned by the *track_request* decorator is the
+    server API function we want to be called by FastAPI.
     """
 
     def decorator(f):

logdetective/server/plot.py

@@ -1,13 +1,18 @@
 import datetime
-from typing import Optional, Union
+from typing import Optional, Union, Dict
 
 import numpy
 import matplotlib
 import matplotlib.figure
 import matplotlib.pyplot
 
+from matplotlib.pyplot import cm
 from logdetective.server import models
-from logdetective.server.database.models import AnalyzeRequestMetrics, EndpointType
+from logdetective.server.database.models import (
+    AnalyzeRequestMetrics,
+    EndpointType,
+    Reactions,
+)
 
 
 class Definition:
@@ -145,13 +150,19 @@ def _add_bar_chart(
     ax.grid(True, alpha=0.3)
 
 
-def _add_line_chart(
-    ax: matplotlib.figure.Axes, timestamps: numpy.array, values: numpy.array, label: str
-) -> None:
+def _add_line_chart(  # pylint: disable=too-many-arguments disable=too-many-positional-arguments
+    ax: matplotlib.figure.Axes,
+    timestamps: numpy.array,
+    values: numpy.array,
+    label: str,
+    color: str = "red",
+    set_label: bool = True,
+):
     """Add a red line chart"""
-    ax.plot(timestamps, values, "r-", linewidth=2, label=label)
-    ax.set_ylabel(label, color="red")
-    ax.tick_params(axis="y", labelcolor="red")
+    ax.plot(timestamps, values, color=color, linestyle="-", linewidth=2, label=label)
+    if set_label:
+        ax.set_ylabel(label, color=color)
+        ax.tick_params(axis="y", labelcolor=color)
 
 
 def requests_per_time(
@@ -249,7 +260,9 @@ def average_time_per_responses(  # pylint: disable=too-many-locals
     )
 
     fig, ax1 = matplotlib.pyplot.subplots(figsize=(12, 6))
-    _add_bar_chart(ax1, plot_def, timestamps, average_time, "average response time (seconds)")
+    _add_bar_chart(
+        ax1, plot_def, timestamps, average_time, "average response time (seconds)"
+    )
 
     responses_average_length = (
         AnalyzeRequestMetrics.get_responses_average_length_in_period(
@@ -279,3 +292,138 @@ def average_time_per_responses(  # pylint: disable=too-many-locals
     matplotlib.pyplot.tight_layout()
 
     return fig
+
+
+def _collect_emoji_data(
+    start_time: datetime.datetime, plot_def: Definition
+) -> Dict[str, Dict[datetime.datetime, int]]:
+    """Collect and organize emoji feedback data
+
+    Counts all emojis given to logdetective comments created since start_time.
+    Collect counts in time accordingly to the plot definition.
+    """
+    reactions = Reactions.get_since(start_time)
+    reactions_values_dict: Dict[str, Dict] = {}
+    for comment_created_at, reaction in reactions:
+        comment_created_at_formatted = comment_created_at.strptime(
+            comment_created_at.strftime(plot_def.time_format), plot_def.time_format
+        )
+        if reaction.reaction_type in reactions_values_dict:
+            reaction_values_dict = reactions_values_dict[reaction.reaction_type]
+            if comment_created_at_formatted in reaction_values_dict:
+                reaction_values_dict[comment_created_at_formatted] += reaction.count
+            else:
+                reaction_values_dict[comment_created_at_formatted] = reaction.count
+        else:
+            reaction_values_dict = {comment_created_at_formatted: reaction.count}
+            reactions_values_dict.update({reaction.reaction_type: reaction_values_dict})
+
+    return reactions_values_dict
+
+
+def _plot_emoji_data(  # pylint: disable=too-many-locals
+    ax: matplotlib.figure.Axes,
+    reactions_values_dict: Dict[str, Dict[datetime.datetime, int]],
+    plot_def: Definition,
+    start_time: datetime.datetime,
+    end_time: datetime.datetime,
+):
+    """Plot each emoji's data on its own axis."""
+    emoji_lines = {}
+    emoji_labels = {}
+
+    # Find global min and max y values to set consistent scale
+    all_counts = []
+    for emoji, dict_counts in reactions_values_dict.items():
+        timestamps, counts = create_time_series_arrays(
+            dict_counts, plot_def, start_time, end_time
+        )
+        all_counts.extend(counts)
+
+    colors = [cm.viridis(i) for i in numpy.linspace(0, 1, len(reactions_values_dict))]    # pylint: disable=no-member
+
+    first_emoji = True
+    for i, (emoji, dict_counts) in enumerate(reactions_values_dict.items()):
+        timestamps, counts = create_time_series_arrays(
+            dict_counts, plot_def, start_time, end_time
+        )
+
+        if first_emoji:
+            current_ax = ax
+            first_emoji = False
+        else:
+            current_ax = ax.twinx()
+            current_ax.spines["right"].set_position(("outward", 60 * (i - 1)))
+
+        _add_line_chart(current_ax, timestamps, counts, f"{emoji}", colors[i], False)
+        emoji_lines[emoji], emoji_labels[emoji] = current_ax.get_legend_handles_labels()
+
+        # Set the same y-limits for all axes
+        current_ax.set_ylim(0, max(all_counts) * 1.1)
+
+        # Only show y-ticks on the first axis to avoid clutter
+        if 0 < i < len(reactions_values_dict):
+            current_ax.set_yticks([])
+
+    return emoji_lines, emoji_labels
+
+
+def emojis_per_time(
+    period_of_time: models.TimePeriod,
+    end_time: Optional[datetime.datetime] = None,
+) -> matplotlib.figure.Figure:
+    """
+    Generate a visualization of overall emoji feedback
+    over a specified time period.
+
+    This function creates a multiple-axis plot showing
+    a line chart for every found emoji
+
+    The time intervals are determined by the provided TimePeriod object, which defines
+    the granularity and formatting of the time axis.
+
+    Args:
+        period_of_time: A TimePeriod object that defines the time period and interval
+                        for the analysis (e.g., hourly, daily, weekly)
+        end_time: The end time for the analysis period. If None, defaults to the current
+                  UTC time
+
+    Returns:
+        A matplotlib Figure object containing the generated visualization
+    """
+    plot_def = Definition(period_of_time)
+    end_time = end_time or datetime.datetime.now(datetime.timezone.utc)
+    start_time = period_of_time.get_period_start_time(end_time)
+    reactions_values_dict = _collect_emoji_data(start_time, plot_def)
+
+    fig, ax = matplotlib.pyplot.subplots(figsize=(12, 6))
+
+    emoji_lines, emoji_labels = _plot_emoji_data(
+        ax, reactions_values_dict, plot_def, start_time, end_time
+    )
+
+    matplotlib.pyplot.title(
+        f"Emoji feedback ({start_time.strftime(plot_def.time_format)} "
+        f"to {end_time.strftime(plot_def.time_format)})"
+    )
+
+    all_lines = []
+    for lines in emoji_lines.values():
+        all_lines.extend(lines)
+    all_labels = []
+    for labels in emoji_labels.values():
+        all_labels.extend(labels)
+
+    ax.legend(all_lines, all_labels, loc="upper left")
+    ax.set_xlabel("Time")
+    ax.set_ylabel("Count")
+
+    # Format x-axis
+    ax.xaxis.set_major_formatter(matplotlib.dates.DateFormatter(plot_def.time_format))
+    ax.xaxis.set_major_locator(plot_def.locator)
+    ax.tick_params(axis="x", labelrotation=45)
+    ax.grid(True, alpha=0.3)
+
+    matplotlib.pyplot.tight_layout()
+
+    return fig

logdetective/server/server.py

@@ -158,8 +158,8 @@ async def analyze_log(
     return Response(explanation=response, response_certainty=certainty)
 
 
-@track_request()
 @app.post("/analyze/staged", response_model=StagedResponse)
+@track_request()
 async def analyze_log_staged(
     build_log: BuildLog, http_session: aiohttp.ClientSession = Depends(get_http_session)
 ):
@@ -420,6 +420,7 @@ class Plot(str, Enum):
 
     REQUESTS = "requests"
     RESPONSES = "responses"
+    EMOJIS = "emojis"
     BOTH = ""
 
 
@@ -450,12 +451,16 @@ async def get_metrics(
                 period_since_now, endpoint_type
             )
             return _svg_figure_response(fig)
+        if plot == Plot.EMOJIS:
+            fig = plot_engine.emojis_per_time(period_since_now)
+            return _svg_figure_response(fig)
         # BOTH
         fig_requests = plot_engine.requests_per_time(period_since_now, endpoint_type)
         fig_responses = plot_engine.average_time_per_responses(
             period_since_now, endpoint_type
         )
-        return _multiple_svg_figures_response([fig_requests, fig_responses])
+        fig_emojis = plot_engine.emojis_per_time(period_since_now)
+        return _multiple_svg_figures_response([fig_requests, fig_responses, fig_emojis])
 
     descriptions = {
         Plot.REQUESTS: (
@@ -466,6 +471,10 @@ async def get_metrics(
             "Show statistics for responses given in the specified period of time "
             f"for the /{endpoint_type.value} API endpoint."
         ),
+        Plot.EMOJIS: (
+            "Show statistics for emoji feedback in the specified period of time "
+            f"for the /{endpoint_type.value} API endpoint."
+        ),
         Plot.BOTH: (
             "Show statistics for requests and responses in the given period of time "
             f"for the /{endpoint_type.value} API endpoint."

{logdetective-0.10.0.dist-info → logdetective-0.11.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: logdetective
-Version: 0.10.0
+Version: 0.11.2
 Summary: Log using LLM AI to search for build/test failures and provide ideas for fixing these.
 License: Apache-2.0
 Author: Jiri Podivin
@@ -48,14 +48,16 @@ Log Detective
 
 [PyPI Releases]: https://pypi.org/project/logdetective/#history
 
-A Python tool to analyze logs using a Language Model (LLM) and Drain template miner.
+A tool, service and RHEL process integration to analyze logs using a Large Language Model (LLM) and a [Drain template miner](https://github.com/logpai/Drain3).
+
+The service that explains logs is available here: https://logdetective.com/explain
 
 Note: if you are looking for code of website logdetective.com it is in [github.com/fedora-copr/logdetective-website](https://github.com/fedora-copr/logdetective-website).
 
 Installation
 ------------
 
-**Fedora 40+**
+**Fedora 41+**
 
     dnf install logdetective
 
@@ -70,11 +72,12 @@ First, ensure that the necessary dependencies for the `llama-cpp-python` project
 
 Then, install the `logdetective` project using pip:
 
-    # then install logdetective project
     pip install logdetective
 
 **Local repository install**
 
+Clone this repository and install with pip:
+
     pip install .
 
 Usage
@@ -111,14 +114,14 @@ Example of altered prompts:
      logdetective https://kojipkgs.fedoraproject.org//work/tasks/3367/131313367/build.log --prompts ~/my-prompts.yml
 
 
-Note that streaming with some models (notably Meta-Llama-3 is broken) is broken and can be workarounded by `no-stream` option:
+Note that streaming with some models (notably Meta-Llama-3 is broken) is broken and can be worked around by `no-stream` option:
 
     logdetective https://example.com/logs.txt --model QuantFactory/Meta-Llama-3-8B-Instruct-GGUF --no-stream
 
 
 Real Example
 ------------
-Let's have a look at a real world example. Log Detective can work with any logs though we optimize it for build logs.
+Let's have a look at a real world example. Log Detective can work with any logs though we optimize it for RPM build logs.
 
 We're going to analyze a failed build of a python-based library that happened in Fedora Koji buildsystem:
 ```
@@ -184,8 +187,13 @@ Contributing
 ------------
 
 Contributions are welcome! Please submit a pull request if you have any improvements or new features to add. Make sure your changes pass all existing tests before submitting.
+For bigger code changes, please consult us first by creating an issue.
+
+We are always looking for more annotated snippets that will increase the quality of Log Detective's results. The contributions happen in our website: https://logdetective.com/
+
+Log Detective performs several inference queries while evaluating a log file. Prompts are stored in a separate file (more info below: https://github.com/fedora-copr/logdetective?tab=readme-ov-file#system-prompts). If you have an idea for improvements to our prompts, please open a PR and we'd happy to test it out.
 
-To develop logdetective, you should fork this repository, clone your fork, and install dependencies using pip:
+To develop Log Detective, you should fork this repository, clone your fork, and install dependencies using pip:
 
     git clone https://github.com/yourusername/logdetective.git
     cd logdetective
@@ -358,7 +366,7 @@ using either a browser, the `curl` or the `http` command (provided by the `httpi
 
 When no time period is specified, the query defaults to the last 2 days:
 
-You can view requests and responses statistics
+You can view requests, responses and emojis statistics
  - for the `/analyze` endpoint at http://localhost:8080/metrics/analyze
  - for the `/analyze-staged` endpoint at http://localhost:8080/metrics/analyze-staged.
  - for the requests coming from gitlab: http://localhost:8080/metrics/analyze-gitlab.
@@ -370,6 +378,7 @@ You can retrieve single svg images at the following endpoints:
  - http://localhost:8080/metrics/analyze-staged/responses
  - http://localhost:8080/metrics/analyze-gitlab/requests
  - http://localhost:8080/metrics/analyze-gitlab/responses
+ - http://localhost:8080/metrics/analyze-gitlab/emojis
 
 Examples:
 

{logdetective-0.10.0.dist-info → logdetective-0.11.2.dist-info}/RECORD

@@ -12,20 +12,20 @@ logdetective/server/config.py,sha256=S2kuvzEo801Kq0vJpRr2fVxSqPghg9kW1L0Ml2yH8Zk
 logdetective/server/database/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 logdetective/server/database/base.py,sha256=1mcjEbhwLl4RalvT3oy6XVctjJoWIW3H9aI_sMWJBK8,1728
 logdetective/server/database/models/__init__.py,sha256=xy2hkygyw6_87zPKkG20i7g7_LXTGR__PUeojhbvv94,496
-logdetective/server/database/models/merge_request_jobs.py,sha256=5WOjJZZYaNMOlegB7Aty96k3NTCutcLP3WW8Yc9sENs,18105
+logdetective/server/database/models/merge_request_jobs.py,sha256=hw88wV1-3x7i53sX7ZotKClc6OsH1njPpbRSZofnqr4,18670
 logdetective/server/database/models/metrics.py,sha256=yl9fS4IPVFWDeFvPAxO6zOVu6oLF319ApvVLAgnD5yU,13928
 logdetective/server/emoji.py,sha256=g9GtMChwznD8g1xonsh-I_3xqRn6LBeg3sjPJWcI0Yg,3333
 logdetective/server/gitlab.py,sha256=fpJp28YsvHvm4DjrvzrgamLk31Fo5UyvT6GNWway9KM,15227
-logdetective/server/llm.py,sha256=AXOqPbx3NJJpf2lOym3EYuVc2Uaf5UUORVQy8Nh2xdM,8516
-logdetective/server/metric.py,sha256=Jf5s_C464VQ2BRlHxq8CjjV7yJ9ZAE0ubnrYF6xr914,3352
+logdetective/server/llm.py,sha256=JtSCZj8SLnoyTCUdhA0TwcsMZfmHFFru2bJ9txI3GuU,8727
+logdetective/server/metric.py,sha256=B3ew_qSmtEMj6xl-FoOtS4F_bkplp-shhtfHF1cG_Io,4010
 logdetective/server/models.py,sha256=eNEB3WJWeZ9Pe6qsmTKQwAE8wu8u51OwLILzV9__YJM,11248
-logdetective/server/plot.py,sha256=B2rOngqx7g-Z3NfttboTip3frkypdF1H7FhK8vh45mE,9655
-logdetective/server/server.py,sha256=7E0x9t1MHICXp6sMgc5Xj3IgbVmYzAvPSJ__brqf5fI,17709
+logdetective/server/plot.py,sha256=eZs4r9gua-nW3yymSMIz1leL9mb4QKlh6FJZSeOfZ5M,14872
+logdetective/server/server.py,sha256=9shFgRkWcJVM2L7HHoQBMCfKuJamh2L4tC96duFPEOA,18127
 logdetective/server/templates/gitlab_full_comment.md.j2,sha256=DQZ2WVFedpuXI6znbHIW4wpF9BmFS8FaUkowh8AnGhE,1627
 logdetective/server/templates/gitlab_short_comment.md.j2,sha256=fzScpayv2vpRLczP_0O0YxtA8rsKvR6gSv4ntNdWb98,1443
 logdetective/utils.py,sha256=hdExAC8FtDIxvdgIq-Ro6LVM-JZ-k_UofaMzaDAHvzM,6088
-logdetective-0.10.0.dist-info/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
-logdetective-0.10.0.dist-info/METADATA,sha256=7ObUbKaJBF-dUWDGZlLlb-I9xGb0se6nSXS5B78vwDk,16340
-logdetective-0.10.0.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-logdetective-0.10.0.dist-info/entry_points.txt,sha256=3K_vXja6PmcA8sNdUi63WdImeiNhVZcEGPTaoJmltfA,63
-logdetective-0.10.0.dist-info/RECORD,,
+logdetective-0.11.2.dist-info/LICENSE,sha256=z8d0m5b2O9McPEK1xHG_dWgUBT6EfBDz6wA0F7xSPTA,11358
+logdetective-0.11.2.dist-info/METADATA,sha256=BVRCWRVzlm-Aa0b51d0ZCbKz2ty1htTzHzX9XGiXALI,17137
+logdetective-0.11.2.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+logdetective-0.11.2.dist-info/entry_points.txt,sha256=3K_vXja6PmcA8sNdUi63WdImeiNhVZcEGPTaoJmltfA,63
+logdetective-0.11.2.dist-info/RECORD,,