Add docstring to _open_for_reading. Remove old comments.

src/biocommons/seqrepo/fastadir/fastadir.py

@@ -139,7 +139,6 @@ def fetch(self, seq_id, start=None, end=None):
         """fetch sequence by seq_id, optionally with start, end bounds
 
         """
-        # INFO path: /Users/kferrite/dev/biocommons.seqrepo/seqrepo/2021-01-29/sequences/2016/0824/050304/1472014984.5124342.fa.bgz, seq_id: v_QTc1p-MUYdgrRv4LMT6ByXIOsdw3C_, start: None, end: None
         rec = self.fetch_seqinfo(seq_id)
 
         if self._writing and self._writing["relpath"] == rec["relpath"]:
@@ -151,14 +150,10 @@ def fetch(self, seq_id, start=None, end=None):
             "path: %s, seq_id: %s, start: %s, end: %s",
             path, seq_id, start, end)
 
-        # fabgz = self._open_for_reading(path)
-        # return fabgz.fetch(seq_id, start, end)
-
         with self._open_for_reading(path) as fabgz:
             seq = fabgz.fetch(seq_id, start, end)
         return seq
 
-
     @functools.lru_cache(maxsize=SEQREPO_LRU_CACHE_MAXSIZE)
     def fetch_seqinfo(self, seq_id):
         """fetch sequence info by seq_id
@@ -241,6 +236,12 @@ def _upgrade_db(self):
 
     @functools.lru_cache()
     def _open_for_reading(self, path):
+        """
+        Opens a FabgzReader to path, wraps in a LockableFabgzReader for use in context managers.
+        Places it in an LRU cache so file is only opened once per FastaDir object. Caller must
+        lock the LockableFabgzReader or otherwise handle concurrent access if sharing between
+        in-process concurrent execution threads, such as asyncio (e.g. WSGI/ASGI web servers)
+        """
         _logger.debug("Opening for reading: %s", path)
         if not os.path.exists(path):
             _logger.error("_open_for_reading path does not exist: %s", path)