Miscellaneous changes

Author: jeromekelleher

examples/1kg_conf.toml

@@ -7,7 +7,7 @@
 name = "1kgp_chr20"
 path = "data/1kgp_chr20.vcz"
 # First 5Mb for development
-regions = "chr20:1-10000000"
+# regions = "chr20:1-10000000"
 include = 'TYPE="snp"'
 
 # wget https://ftp.ensembl.org/pub/release-115/variation/vcf/homo_sapiens/homo_sapiens-chr20.vcf.gz
@@ -24,9 +24,7 @@ name = "ancestors"
 path = "data/ancestors.vcz"          # output path for the ancestor store
 sources = ["1kgp_chr20"]             # which source(s) to build from
 max_gap_length = 500_000
-genotype_encoding = "one_bit"
-# samples_chunk_size  = 1000         # default: 1000 (ancestor dimension)
-# variants_chunk_size = 1000         # default: 1000 (site dimension)
+samples_chunk_size  = 100         
 
 [match]
 output = "data/output.trees"         # output tree sequence path

tsinfer/config.py

@@ -41,8 +41,8 @@
 # Default constants
 # ---------------------------------------------------------------------------
 DEFAULT_MAX_GAP_LENGTH = 500_000
-DEFAULT_SAMPLES_CHUNK_SIZE = 1000
-DEFAULT_VARIANTS_CHUNK_SIZE = 1000
+DEFAULT_SAMPLES_CHUNK_SIZE = 100
+DEFAULT_VARIANTS_CHUNK_SIZE = 50_000
 DEFAULT_GENOTYPE_ENCODING = 0  # 0 = eight-bit, 1 = one-bit
 DEFAULT_COMPRESSOR = "zstd"
 DEFAULT_COMPRESSION_LEVEL = 7

tsinfer/pipeline.py

@@ -502,8 +502,8 @@ def match(
         if progress:
             pbar = tqdm.tqdm(
                 total=len(job_list),
-                desc=f"Group {gi + 1}/{num_groups}",
-                unit="haplotypes",
+                desc=f"Group {gi} ({num_groups})",
+                unit="haps",
             )
         results = []
         for job, result in match_iter:

tsinfer/vcz.py

@@ -48,6 +48,7 @@
 import queue
 import threading
 import time as _time
+import warnings
 from typing import Any
 
 import numpy as np
@@ -2088,8 +2089,6 @@ def __init__(
                 chunks_fit,
             )
             if cb > 0 and chunks_fit < 2:
-                import warnings
-
                 warnings.warn(
                     f"Cache can hold fewer than 2 chunks from source "
                     f"'{name}' (chunk={cb / (1024 * 1024):.1f} MiB, "