philiprehberger-import-profiler
v0.3.0995 downloadsMIT
Show how long each Python import takes during startup
$
pip install philiprehberger-import-profilerREADME
philiprehberger-import-profiler
Show how long each Python import takes during startup.
Installation
pip install philiprehberger-import-profiler
Usage
Basic Profiling
from philiprehberger_import_profiler import profile_imports
report = profile_imports("my_package")
# Slowest imports
for entry in report.slowest(10):
print(f"{entry.name}: {entry.duration_ms:.1f}ms")
# Tree view
report.print_tree(threshold_ms=5.0)
# └── my_package (124.5ms)
# ├── requests (45.2ms)
# │ └── urllib3 (22.1ms)
# └── numpy (62.0ms)
# Summary
print(f"Total: {report.total_ms:.1f}ms, Modules: {report.module_count}")
Self Time and Export
report = profile_imports("my_package")
# Self time excludes children's duration
for entry in report.slowest(5):
print(f"{entry.name}: total={entry.duration_ms:.1f}ms self={entry.self_ms:.1f}ms")
# Export as list of dicts
data = report.to_dict()
# [{"name": "requests", "duration_ms": 45.2, "self_ms": 23.1, "parent": "my_package"}, ...]
Summary and JSON export
Use summary() for a compact dict you can log or pretty-print, and to_json() for the full entry list as a JSON string.
report = profile_imports("my_package")
print(report.summary(slowest=3))
# {"total_ms": 124.5, "module_count": 42, "slowest": [
# {"name": "numpy", "duration_ms": 62.0},
# {"name": "requests", "duration_ms": 45.2},
# {"name": "urllib3", "duration_ms": 22.1},
# ]}
# JSON string (compact when indent=None)
print(report.to_json(indent=None))
Sorting and filtering reports
report = profile_imports("my_package")
# Top N slowest entries (descending by duration_ms)
for entry in report.slowest(3):
print(f"{entry.name}: {entry.duration_ms:.1f}ms")
# Filter to entries whose module name starts with a prefix
numpy_only = report.filter("numpy")
print(f"numpy.* modules: {numpy_only.module_count}, total={numpy_only.total_ms:.1f}ms")
# filter returns a new report; the original is unchanged
assert report.module_count >= numpy_only.module_count
API
| Function / Class | Description |
|---|---|
profile_imports(module_name) | Profile all imports, returns ImportReport |
report.slowest(n=10) | Top N slowest imports as ImportEntry list, sorted descending |
report.filter(prefix) | New ImportReport containing only entries whose name starts with prefix |
report.print_tree(threshold_ms=0) | Print indented tree |
report.total_ms | Total import time in milliseconds |
report.module_count | Number of modules imported |
report.to_dict() | Export as list of dicts |
report.to_json(indent=2) | Export entries as a JSON-formatted string |
report.summary(slowest=5) | Compact dict with total_ms, module_count, and top-N slowest |
ImportEntry.name | Module name |
ImportEntry.duration_ms | Total duration including children |
ImportEntry.self_ms | Duration excluding children |
Development
pip install -e .
python -m pytest tests/ -v
Support
If you find this project useful: