Moving loading and save sigs to rust

move json parsing and init to rust
working on loading sigs

55 failing. Now it's failing because SBT index is saving all signatures
(instead of only the one it was used to build the tree).
This was actually a feature (see #198) but it broke the SBT code
(it wasn't ready for that!)

setup.py

@@ -60,8 +60,11 @@ def build_native(spec):
     "packages": find_packages(),
     "zip_safe": False,
     "platforms": "any",
-    "entry_points": {"console_scripts": ["sourmash = sourmash.__main__:main"]},
-    "install_requires": ["screed>=0.9", "ijson", "khmer>=2.1", "milksnake"],
+    "entry_points": {'console_scripts': [
+        'sourmash = sourmash.__main__:main'
+        ]
+    },
+    "install_requires": ["screed>=0.9", "khmer>=2.1", 'milksnake'],
     "setup_requires": [
         "setuptools>=38.6.0",
         "milksnake",

sourmash/_compat.py

@@ -22,3 +22,12 @@ def implements_to_string(cls):
     itervalues = lambda x: x.values()
     NUL = 0
     implements_to_string = lambda x: x
+
+
+def to_bytes(s):
+    if not isinstance(s, string_types + (bytes,)):
+        raise TypeError("Requires a string-like sequence")
+
+    if isinstance(s, string_types):
+        s = s.encode('utf-8')
+    return s

sourmash/_minhash.py

@@ -4,8 +4,8 @@
 import math
 import copy
 
-from ._compat import string_types, range_type
 from ._lowlevel import ffi, lib
+from ._compat import to_bytes
 from .utils import RustObject, rustcall
 
 # default MurmurHash seed
@@ -40,15 +40,6 @@ def get_scaled_for_max_hash(max_hash):
     return int(round(get_minhash_max_hash() / max_hash, 0))
 
 
-def to_bytes(s):
-    if not isinstance(s, string_types + (bytes,)):
-        raise TypeError("Requires a string-like sequence")
-
-    if isinstance(s, string_types):
-        s = s.encode("utf-8")
-    return s
-
-
 def hash_murmur(kmer, seed=MINHASH_DEFAULT_SEED):
     "hash_murmur(string, [,seed])\n\n"
     "Compute a hash for a string, optionally using a seed (an integer). "
@@ -92,8 +83,6 @@ def __init__(
         mins=None,
         scaled=0,
     ):
-        self.track_abundance = track_abundance
-
         if max_hash and scaled:
             raise ValueError("cannot set both max_hash and scaled")
         elif scaled:
@@ -221,6 +210,10 @@ def subtract_mins(self, other):
         b = set(other.get_mins())
         return a - b
 
+    @property
+    def track_abundance(self):
+        return self._methodcall(lib.kmerminhash_track_abundance)
+
     @property
     def seed(self):
         return self._methodcall(lib.kmerminhash_seed)

sourmash/commands.py

@@ -1325,14 +1325,14 @@ def watch(args):
         ksize = tree_mh.ksize
 
     E = MinHash(ksize=ksize, n=args.num_hashes, is_protein=is_protein)
-    streamsig = sig.SourmashSignature(E, filename='stdin', name=args.name)
 
     notify('Computing signature for k={}, {} from stdin', ksize, moltype)
 
     def do_search():
         search_fn = SearchMinHashesFindBest().search
 
         results = []
+        streamsig = sig.SourmashSignature(E, filename='stdin', name=args.name)
         for leaf in tree.find(search_fn, streamsig, args.threshold):
             results.append((streamsig.similarity(leaf.data),
                             leaf.data))
@@ -1370,6 +1370,7 @@ def do_search():
 
     if args.output:
         notify('saving signature to {}', args.output.name)
+        streamsig = sig.SourmashSignature(E, filename='stdin', name=args.name)
         sig.save_signatures([streamsig], args.output)