Fixes #16: Detecting EOF for serial access

The program was originally looking for the EOF marker, but I forgot that when the EOF marker BGZF block is decompressed, it evaluates to an empty bytestring `b''`. I put in some extra checks so that if an empty block is found, it ensures it is at the file and is a legitimate end of file.

If the file is truncated, it will `raise` a slightly different `StopIteration` error stating that the *potential* end of file is reached.

I ran this on the file you mentioned in #15 (btw, awesome Issue suggestion. Thank you for following the template). My debugging is properly running through the file now.

bamnostic/bgzf.py

@@ -453,9 +453,11 @@ def __init__(self, filepath_or_object, mode="rb", max_cache=128, index_filename=
         # Connect to the BAM file
         self._handle = handle
 
+        self._truncated = self._check_truncation()
+        self._igore_truncation = ignore_truncation
         # Check BAM file integrity
-        if not ignore_truncation:
-            if self._check_truncation():
+        if not self._igore_truncation:
+            if self._truncated:
                 raise Exception('BAM file may be truncated. Turn off ignore_truncation if you wish to continue')
 
         # Connect and process the Index file (if present)
@@ -525,7 +527,7 @@ def _load_block(self, start_offset=None):
             if self._text:
                 self._buffer = ""
             else:
-                self._buffer = b""
+                self._buffer = b''
         self._within_block_offset = 0
         self._block_raw_length = block_size
 

bamnostic/core.py

@@ -14,7 +14,7 @@
 @email: "mdsherman<at>betteridiot<dot>tech"
 
 """
-
+import os
 import struct
 import sys
 from array import array
@@ -167,7 +167,20 @@ def __init__(self, _io):
 
         """
         self._io = _io
-        block_size = unpack_int32(self._io.read(4))[0]
+        bsize_buffer = self._io.read(4)
+        try:
+            block_size = unpack_int32(bsize_buffer)[0]
+
+        # Check for EOF: If the cursor is at the end of file, read() will return 
+        # an empty byte string. If 
+        except struct.error:
+            if all([not bsize_buffer, not self._io._handle.read()]):
+                if not self._io._igore_truncation and not self._io._truncated:
+                    raise StopIteration('End of file reached')
+                else:
+                    raise StopIteration('Potential end of file reached')
+            else:
+                raise IOError('Reached End of file, but marker does not match BAM standard')
 
         # Pull in the whole read
         self._byte_stream = bytearray(self._io.read(block_size))

codemeta.json

@@ -37,8 +37,8 @@
   "description": "BAMnostic is a Pure Python OS, version, and runtime agnostic BAM file parser",
   "keywords": "BAM, pysam, genomics, genetics, htslib, samtools",
   "license": "https://github.com/betteridiot/bamnostic/blob/master/LICENSE",
-  "softwareVersion": "v0.8.13",
-  "version": "v0.8.13",
+  "softwareVersion": "v0.9.0",
+  "version": "v0.9.0",
   "readme": "https://github.com/betteridiot/bamnostic/blob/master/README.md",
   "buildInstructions": "https://github.com/betteridiot/bamnostic/blob/master/README.md",
   "issueTracker": "https://github.com/betteridiot/bamnostic/issues",

docs/paper/codemeta.json

@@ -37,8 +37,8 @@
   "description": "BAMnostic: an OS-agnostic toolkit for genomic sequence analysis",
   "keywords": "BAM, pysam, genomics, genetics, htslib, samtools",
   "license": "https://github.com/betteridiot/bamnostic/blob/master/LICENSE",
-  "softwareVersion": "v0.8.13",
-  "version": "v0.8.13",
+  "softwareVersion": "v0.9.0",
+  "version": "v0.9.0",
   "readme": "https://github.com/betteridiot/bamnostic/blob/master/README.md",
   "buildInstructions": "https://github.com/betteridiot/bamnostic/blob/master/README.md",
   "issueTracker": "https://github.com/betteridiot/bamnostic/issues",

setup.py

@@ -9,7 +9,7 @@ def readme():
 
 setup(
     name='bamnostic',
-    version='0.8.13',
+    version='0.9.0',
     description='Pure Python, OS-agnostic Binary Alignment Map (BAM) random access and parsing tool',
     long_description=readme(),
     url='https://github.com/betteridiot/bamnostic/',