TableLogger

Common errors

• ImportError: numpy.core.multiarray failed to import

  cause This error, particularly after a NumPy upgrade to 2.0+, often indicates an ABI (Application Binary Interface) incompatibility. Python packages that use NumPy's C API and were built against an older NumPy version will not work with NumPy 2.0.
  fix Reinstall `table-logger` and any other packages that depend on NumPy after upgrading NumPy. This ensures they are compiled against the new NumPy ABI. Example: `pip uninstall numpy table-logger && pip install numpy table-logger`.

• Table output not displaying correctly in Jupyter notebooks.

  cause Older versions of `table-logger` had specific issues handling output streams in Jupyter environments.
  fix Update to `table-logger` version 0.3.5 or newer: `pip install --upgrade table-logger`.

• TypeError: 'numpy.float64' object is not callable or AttributeError: module 'numpy' has no attribute 'float'

  cause This typically occurs if your `table-logger` version is older and attempts to use deprecated NumPy aliases like `np.float` or `np.int`, which have been removed in newer NumPy versions. Version 0.3.7 of `table-logger` explicitly removed its internal use of these.
  fix Upgrade `table-logger` to 0.3.7 (once available on PyPI) or ensure your code is using modern NumPy types (e.g., `float` or `int` directly) and that all dependent libraries are compatible with your NumPy version.

Warnings

• breaking Version 0.3.7 (GitHub release) removed deprecated `numpy` aliases `np.float` and `np.int`. If you are using an older version of `table-logger` with a very new `numpy` version (e.g., NumPy 2.0+), or your code relies on these aliases in conjunction with `table-logger`, you may encounter errors related to these removed types.
  Fix: Upgrade `table-logger` to 0.3.7 or newer once available on PyPI, and ensure your code adapts to modern NumPy type handling (e.g., using `float` or `int` directly where appropriate). Ensure all `numpy` dependent packages are compatible with your `numpy` version.
• gotcha Prior to version 0.3.5, users reported logging issues when `table-logger` was used within Jupyter notebooks, where output might not display correctly or consistently.
  Fix: Upgrade to `table-logger` version 0.3.5 or newer to resolve known Jupyter notebook logging inconsistencies.
• gotcha NumPy 2.0 (released June 2024) introduced significant breaking changes, including an ABI break and API cleanups. If `table-logger` (or other libraries it depends on) was compiled against an older NumPy 1.x version and then run with NumPy 2.0, this can lead to `ImportError` due to binary incompatibility.
  Fix: Reinstall `table-logger` and its dependencies after upgrading NumPy to ensure they are compiled against the correct NumPy version. Check for `table-logger` releases that explicitly support NumPy 2.0.

Install

• pip install table-logger Install with pip

Imports

• TableLogger

  from table_logger import TableLogger

Quickstart

Initialize a TableLogger instance with specified columns and then log rows. The `flush()` method is implicitly called when the object is garbage collected or the program exits, but can be called explicitly to ensure output. Examples also show enabling row numbers, timestamps, and time deltas.

from table_logger import TableLogger
from datetime import datetime
import math
import random

tbl = TableLogger(columns='a,b,c,d')
tbl(1, 'Row1', datetime.now(), math.pi)
tbl(2, 'Row2', datetime.now(), 1/3)
tbl(3, 'Row3', datetime.now(), random.random())

# Example with additional columns
import time
tbl_extended = TableLogger(columns='data', rownum=True, time_delta=True, timestamp=True)
for e in 'abcde':
    time.sleep(random.uniform(0.1, 0.5)) # Use uniform for more realistic varying delays
    tbl_extended(e)

print('\nStandard Table:')
tbl.flush()
print('\nExtended Table:')
tbl_extended.flush()