Set encoding="utf-8" to avoid Windows default encoding issue

Author: moshi4

src/pycirclize/parser/bed.py

@@ -52,7 +52,7 @@ def parse(bed_file: str | Path) -> list[BedRecord]:
             BED records
         """
         bed_records = []
-        with open(bed_file) as f:
+        with open(bed_file, encoding="utf-8") as f:
             reader = csv.reader(f, delimiter="\t")
             for row in reader:
                 if row[0].startswith("#") or len(row) < 3:

src/pycirclize/parser/genbank.py

@@ -383,7 +383,7 @@ def write_genome_fasta(self, outfile: str | Path) -> None:
         outfile : str | Path
             Output genome fasta file
         """
-        with open(outfile, "w") as f:
+        with open(outfile, "w", encoding="utf-8") as f:
             for seqid, seq in self.get_seqid2seq().items():
                 f.write(f">{seqid}\n{seq}\n")
 
@@ -406,19 +406,23 @@ def _parse_gbk_source(
         list[SeqRecord]
             Genbank SeqRecords
         """
-        # Parse compressed file
+        # Parse file
         if isinstance(gbk_source, (str, Path)):
             if Path(gbk_source).suffix == ".gz":
-                with gzip.open(gbk_source, mode="rt") as f:
+                with gzip.open(gbk_source, mode="rt", encoding="utf-8") as f:
                     return list(SeqIO.parse(f, "genbank"))
             elif Path(gbk_source).suffix == ".bz2":
-                with bz2.open(gbk_source, mode="rt") as f:
+                with bz2.open(gbk_source, mode="rt", encoding="utf-8") as f:
                     return list(SeqIO.parse(f, "genbank"))
             elif Path(gbk_source).suffix == ".zip":
                 with zipfile.ZipFile(gbk_source) as zip:
                     with zip.open(zip.namelist()[0]) as f:
-                        return list(SeqIO.parse(TextIOWrapper(f), "genbank"))
-        # Parse no compressed file or TextIOWrapper
+                        io = TextIOWrapper(f, encoding="utf-8")
+                        return list(SeqIO.parse(io, "genbank"))
+            else:
+                with open(gbk_source, encoding="utf-8") as f:
+                    return list(SeqIO.parse(f, "genbank"))
+        # Parse TextIOWrapper
         return list(SeqIO.parse(gbk_source, "genbank"))
 
     def _is_straddle_feature(self, feature: SeqFeature) -> bool:

src/pycirclize/parser/gff.py

@@ -280,18 +280,18 @@ def _parse_gff(
         """
         gff_file = Path(gff_file)
         if gff_file.suffix == ".gz":
-            with gzip.open(gff_file, mode="rt") as f:
+            with gzip.open(gff_file, mode="rt", encoding="utf-8") as f:
                 gff_records, start, end = self._parse_gff_textio(f, target_seqid)
         elif gff_file.suffix == ".bz2":
-            with bz2.open(gff_file, mode="rt") as f:
+            with bz2.open(gff_file, mode="rt", encoding="utf-8") as f:
                 gff_records, start, end = self._parse_gff_textio(f, target_seqid)
         elif gff_file.suffix == ".zip":
             with zipfile.ZipFile(gff_file) as zip:
                 with zip.open(zip.namelist()[0]) as f:
-                    io = TextIOWrapper(f)
+                    io = TextIOWrapper(f, encoding="utf-8")
                     gff_records, start, end = self._parse_gff_textio(io, target_seqid)
         else:
-            with open(gff_file) as f:
+            with open(gff_file, encoding="utf-8") as f:
                 gff_records, start, end = self._parse_gff_textio(f, target_seqid)
 
         return gff_records, start, end

src/pycirclize/tree.py

@@ -184,10 +184,12 @@ def load_tree(data: str | Path | Tree, format: str) -> Tree:
         """
         if isinstance(data, str) and urlparse(data).scheme in ("http", "https"):
             # Load tree file from URL
-            return Phylo.read(io.StringIO(urlopen(data).read().decode()), format=format)
+            treeio = io.StringIO(urlopen(data).read().decode(encoding="utf-8"))
+            return Phylo.read(treeio, format=format)
         elif isinstance(data, (str, Path)) and os.path.isfile(data):
             # Load tree file
-            return Phylo.read(data, format=format)
+            with open(data, encoding="utf-8") as f:
+                return Phylo.read(f, format=format)
         elif isinstance(data, str):
             # Load tree string
             return Phylo.read(io.StringIO(data), format=format)

src/pycirclize/utils/dataset.py

@@ -225,7 +225,7 @@ def fetch_genbank_by_accid(
     )
     if gbk_outfile is not None:
         gbk_text = gbk_fetch_data.read()
-        with open(gbk_outfile, "w") as f:
+        with open(gbk_outfile, "w", encoding="utf-8") as f:
             f.write(gbk_text)
         gbk_fetch_data = StringIO(gbk_text)
 
@@ -258,7 +258,7 @@ def load(chr_link_file: str | Path) -> list[ChrLink]:
             Chromosome link list
         """
         chr_link_list = []
-        with open(chr_link_file) as f:
+        with open(chr_link_file, encoding="utf-8") as f:
             reader = csv.reader(f, delimiter="\t")
             for row in reader:
                 qchr, qstart, qend = row[0], int(row[1]), int(row[2])