feat: support specifying circular contigs

bento-platform · Dec 22, 2023 · 18a4159 · 18a4159
1 parent 247c5d2
commit 18a4159
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 3 deletions.
diff --git a/fasta_checksum_utils/entry.py b/fasta_checksum_utils/entry.py
@@ -17,13 +17,14 @@ async def main():
     parser.add_argument("fasta", type=str, help="A FASTA path or URI to checksum.")
     parser.add_argument("--fai", type=str, help="A FASTA FAI index path or URI, if available.")
     parser.add_argument("--genome-id", type=str, help="Genome ID to include, if --out-format is set to bento-json.")
+    parser.add_argument("--circular-contigs", type=str, nargs="*", help="Names of circular contigs in this genome.")
     parser.add_argument(
         "--out-format", type=str, default="text", choices=("text", "bento-json"),
         help="Output format for checksum report; either 'text' or 'bento-json' (default: 'text').")
 
     args = parser.parse_args()
 
-    report = await fasta_report(args.fasta, args.fai, (AlgorithmMD5, AlgorithmGA4GH))
+    report = await fasta_report(args.fasta, args.fai, frozenset(args.circular_contigs), (AlgorithmMD5, AlgorithmGA4GH))
     if args.out_format == "bento-json":
         print(report.as_bento_json(genome_id=getattr(args, "genome_id", None)))
     else:

diff --git a/fasta_checksum_utils/fasta.py b/fasta_checksum_utils/fasta.py
@@ -26,12 +26,14 @@ def __init__(
         file_checksums: dict[ChecksumAlgorithm, str],
         file_size: int,
         sequence_checksums_and_lengths: dict[str, tuple[dict[ChecksumAlgorithm, str], int]],
+        circular_contigs: frozenset[str],
     ):
         self._fasta_path_or_uri: str = fasta_path_or_uri
         self._fai_path_or_uri: Union[str, None] = fai_path_or_uri
         self._file_checksums = file_checksums
         self._file_size: int = file_size
         self._sequence_checksums_and_lengths = sequence_checksums_and_lengths
+        self._circular_contigs: frozenset[str] = circular_contigs
 
     @property
     def fasta_path_or_uri(self) -> str:
@@ -54,8 +56,10 @@ def _checksum_dict(cs: dict[ChecksumAlgorithm, str]) -> dict[str, str]:
             "contigs": [
                 {
                     "name": contig,
+                    "aliases": [],
                     **_checksum_dict(checksums),
-                    "length": length
+                    "length": length,
+                    "circular": contig in self._circular_contigs,
                 }
                 for contig, (checksums, length) in self._sequence_checksums_and_lengths.items()
             ]
@@ -103,6 +107,7 @@ def _is_http_url(x: str) -> bool:
 async def fasta_report(
     fasta_path_or_uri: Union[Path, str],
     fai_path_or_uri: Union[Path, str, None],
+    circular_contigs: frozenset[str],
     algorithms: tuple[ChecksumAlgorithm, ...],
 ) -> FastaReport:
     tmp_file_fa = None
@@ -181,4 +186,4 @@ async def fasta_report(
             os.unlink(tmp_file_fai.name)
 
     # Generate and return a final report
-    return FastaReport(fasta_str, fai_str, file_checksums, file_size, sequence_checksums_and_lengths)
+    return FastaReport(fasta_str, fai_str, file_checksums, file_size, sequence_checksums_and_lengths, circular_contigs)