Refactor package structure

bio2zarr/cli.py

@@ -8,7 +8,7 @@
 import numcodecs
 import tabulate
 
-from . import plink, provenance, vcf, vcf_utils
+from . import icf, plink, provenance, vcf, vcf_utils
 
 logger = logging.getLogger(__name__)
 
@@ -167,7 +167,7 @@ def check_overwrite_dir(path, force):
 def get_compressor(cname):
     if cname is None:
         return None
-    config = vcf.ICF_DEFAULT_COMPRESSOR.get_config()
+    config = icf.ICF_DEFAULT_COMPRESSOR.get_config()
     config["cname"] = cname
     return numcodecs.get_codec(config)
 
@@ -198,7 +198,7 @@ def explode(
     """
     setup_logging(verbose)
     check_overwrite_dir(icf_path, force)
-    vcf.explode(
+    icf.explode(
         icf_path,
         vcfs,
         worker_processes=worker_processes,
@@ -235,7 +235,7 @@ def dexplode_init(
     """
     setup_logging(verbose)
     check_overwrite_dir(icf_path, force)
-    work_summary = vcf.explode_init(
+    work_summary = icf.explode_init(
         icf_path,
         vcfs,
         target_num_partitions=num_partitions,
@@ -263,7 +263,7 @@ def dexplode_partition(icf_path, partition, verbose, one_based):
     setup_logging(verbose)
     if one_based:
         partition -= 1
-    vcf.explode_partition(icf_path, partition)
+    icf.explode_partition(icf_path, partition)
 
 
 @click.command
@@ -274,7 +274,7 @@ def dexplode_finalise(icf_path, verbose):
     Final step for distributed conversion of VCF(s) to intermediate columnar format.
     """
     setup_logging(verbose)
-    vcf.explode_finalise(icf_path)
+    icf.explode_finalise(icf_path)
 
 
 @click.command

bio2zarr/constants.py

@@ -0,0 +1,18 @@
+import numpy as np
+
+INT_MISSING = -1
+INT_FILL = -2
+STR_MISSING = "."
+STR_FILL = ""
+
+FLOAT32_MISSING, FLOAT32_FILL = np.array([0x7F800001, 0x7F800002], dtype=np.int32).view(
+    np.float32
+)
+FLOAT32_MISSING_AS_INT32, FLOAT32_FILL_AS_INT32 = np.array(
+    [0x7F800001, 0x7F800002], dtype=np.int32
+)
+
+
+MIN_INT_VALUE = np.iinfo(np.int32).min + 2
+VCF_INT_MISSING = np.iinfo(np.int32).min
+VCF_INT_FILL = np.iinfo(np.int32).min + 1

bio2zarr/core.py

@@ -1,13 +1,16 @@
 import concurrent.futures as cf
 import contextlib
 import dataclasses
+import json
 import logging
+import math
 import multiprocessing
 import os
 import os.path
 import threading
 import time
 
+import humanfriendly
 import numcodecs
 import numpy as np
 import tqdm
@@ -18,6 +21,17 @@
 numcodecs.blosc.use_threads = False
 
 
+def display_number(x):
+    ret = "n/a"
+    if math.isfinite(x):
+        ret = f"{x: 0.2g}"
+    return ret
+
+
+def display_size(n):
+    return humanfriendly.format_size(n, binary=True)
+
+
 def min_int_dtype(min_value, max_value):
     if min_value > max_value:
         raise ValueError("min_value must be <= max_value")
@@ -277,3 +291,11 @@ def __exit__(self, exc_type, exc_val, exc_tb):
         self._update_progress()
         self.progress_bar.close()
         return False
+
+
+class JsonDataclass:
+    def asdict(self):
+        return dataclasses.asdict(self)
+
+    def asjson(self):
+        return json.dumps(self.asdict(), indent=4)