Skip to content

scripts::extract_source_niches

More...

Functions

Name
extract_source_based_niches(sample_size sample_size =50000)

Attributes

Name
PROJECT_ROOT
exist_ok
dict TARGET_NICHES
viable_niches
source_counts
all_niche_samples
sample_size
dict output_data
output_path
f
indent

Detailed Description

Source-based niche extraction from Common Pile
More reliable than clustering for creating distinct specialists

Functions Documentation

function extract_source_based_niches

extract_source_based_niches(
    sample_size sample_size =50000
)
Extract niches directly from source labels```



## Attributes Documentation

### variable PROJECT_ROOT

```python
PROJECT_ROOT =  Path(__file__).resolve().parent.parent.parent;

variable exist_ok

exist_ok;

variable TARGET_NICHES

dict TARGET_NICHES;

variable viable_niches

viable_niches;

variable source_counts

source_counts;

variable all_niche_samples

all_niche_samples;

variable sample_size

sample_size;

variable output_data

dict output_data =  {
        'viable_niches': viable_niches,
        'all_sources': source_counts,
        'extraction_config': {
            'sample_size': 50000,
            'target_niches': TARGET_NICHES
        }
    };

variable output_path

output_path =  str(PROJECT_ROOT / 'data' / 'analysis' / 'source_based_niches.json');

variable f

f;

variable indent

indent;

Updated on 2026-07-25 at 22:56:57 +0000