pycsvlogger 0.1.1

Flexible, general-purpose CSV-based structured logging for Python

pycsvlogger

A flexible, general-purpose CSV-based structured logging library for Python’s built-in logging framework. It lets you emit rows of CSV with custom columns and automatically captured metadata—so you can track events (downloads, uploads, processing steps, etc.) across your entire codebase in one place.

---

🔧 Features

• Configurable Columns: Define exactly which CSV headers you want (e.g., timestamp, script, record_id, action, status, details, or any custom fields).
• Automatic Metadata: Captures log level (status), message (details), timestamp, and source script filename without extra code.
• Context Binding: Use a LoggerAdapter to bind default fields (e.g. record_id, action, user) once, then simply call .info(), .error(), etc., without repeating metadata.
• No External Dependencies: Built using only Python’s standard library (logging, csv, pathlib).
• High Performance: Appends rows to a single file—no per-record file sprawl.
• Easy Analysis: Load your CSV in Excel, Google Sheets, or use Pandas to filter, pivot, or visualize your logs.

---

🚀 Installation

Install from PyPI:

pip install pycsvlogger

---

🎯 Quickstart

1. Initialize the logger at your application’s entry point, choosing your CSV file and columns:

   from csv_logger import init_csv_logger, get_record_logger
   
   init_csv_logger(
       log_file   = "Logs.csv",
       fieldnames = [
           "timestamp", "script", "record_id", "action", "status", "details"
       ]
   )
   

2. Bind a logger for each unit of work (e.g. per record or per processing step):

   record_id = "abc123"
   logger    = get_record_logger(record_id=record_id, action="download")
   
   logger.info("Starting download from external API")
   # … perform download …
   logger.info("Download complete, saved to transcript.pdf")
   

3. Switch Context when you move to a new action or record:

   # Same record, but now uploading link back
   logger = get_record_logger(record_id=record_id, action="upload_link")
   logger.info("Pushing download URL to record in service")
   # … perform PATCH …
   logger.info("Link updated successfully")
   

Your CSV (logs/app_events.csv) will look like:

timestamp,script,record_id,action,status,details
2025-05-24 11:00:05,main.py,abc123,download,INFO,Starting download from external API
2025-05-24 11:00:23,main.py,abc123,download,INFO,Download complete, saved to transcript.pdf
2025-05-24 11:01:10,main.py,abc123,upload_link,INFO,Pushing download URL to record in service

---

⚙️ Configuration Options

When you call init_csv_logger, you can customize:

• log_file: Path to your CSV log (default: events.csv).

• fieldnames: List of column headers in the order you want.

  • Supported automatic columns:

    • timestamp – formatted as YYYY-MM-DD HH:MM:SS
    • status – the log level (INFO, WARNING, ERROR, etc.)
    • details – your log message
    • script – the base filename of the calling module

  • Custom columns: any header names beyond the built‑in ones (timestamp, status, details, script) will be pulled from the keyword arguments you pass to get_record_logger (for example, record_id="...", action="...", user="..."). just pass all your headers like fieldnames = [

    "timestamp", "script", "record_id", "action", "status", "details"

    ], default + custom. 

• Once set, you don’t need to pass record_id or action on every log call—just bind them once with get_record_logger:

# Bind multiple defaults at once:
logger = get_record_logger(record_id=rid, action="process", user="alice")
logger.info("Step 1: validation complete")
logger.info("Step 2: transformation started")

Those fields (record_id, action, user) will appear as columns in every row.

But whenever you have to assign a new recordid or action just pass them again. You can follow your own headers like username, employeeid e.t.c

---

🔍 Inspecting & Filtering Logs

Since your logs live in a single CSV file, you can easily:

• Excel / Google Sheets: Open and apply a filter on the record_id or action column.

• Command line:

  grep "abc123" logs/app_events.csv
  

• Python / Pandas:

  import pandas as pd
  df = pd.read_csv("logs/app_events.csv")
  # Show only events for record abc123:
  print(df[df.record_id == "abc123"])
  
  # Count how many downloads succeeded:
  print((df[df.action == "download"].status == "INFO").sum())
  

This makes it trivial to drill down on failures, measure performance, or audit the full history for any ID.

---

💡 Advanced Tips

• Error-level Logging: When something goes wrong or you want to highlight warnings, use logger.error("message") or logger.warning("message"). These calls emit rows where the status column equals ERROR or WARNING, respectively, making it trivial to filter your CSV for all error events:

  logger = get_record_logger(record_id=rid, action="download")
  try:
      download_file()
  except Exception as e:
      logger.error(f"Download failed: {e}")
  

• Custom Metadata: Bind any extra contextual information—such as module, session_id, job_name, or business-specific keys like user_id or order_id—by passing them as keyword arguments to get_record_logger. Ensure your fieldnames list includes these column names:

  init_csv_logger(
      log_file="logs/events.csv",
      fieldnames=["timestamp","script","status","details","module","session_id"]
  )
  logger = get_record_logger(session_id="abc123", module="uploader")
  logger.info("Upload started")
  

• Multiple Loggers: If you need separate CSV files for different workflows, simply call init_csv_logger with different log_file paths early in each module.

Happy logging!