update readme (#16)

* get rid of separate include/ dir

* update readme

* version bump

Author: escherba

MANIFEST.in

@@ -1,7 +1,7 @@
 include LICENSE
 include README.rst
 include cpuinfo.py
-include include/*.h
+include src/*.h
 include src/*.cc
 include src/*.cpp
 include src/*.pyx

README.rst

@@ -60,10 +60,12 @@ MetroHash algorithm. For stateless hashing, it exports ``metrohash64`` and
 Incremental hashing
 ~~~~~~~~~~~~~~~~~~~
 
-For incremental hashing, use ``MetroHash64`` and ``MetroHash128`` classes.
-Incremental hashing is associative and guarantees that any combination of input
-slices will result in the same final hash value. This is useful for processing
-large inputs and stream data. Example with two slices:
+Unlike its cousins CityHash and FarmHash, MetroHash allows incremental
+(stateful) hashing.  For incremental hashing, use ``MetroHash64`` and
+``MetroHash128`` classes.  Incremental hashing is associative and guarantees
+that any combination of input slices will result in the same final hash value.
+This is useful for processing large inputs and stream data. Example with two
+slices:
 
 .. code-block:: python
 
@@ -85,10 +87,14 @@ Note that the resulting hash value above is the same as in:
 Buffer protocol support
 ~~~~~~~~~~~~~~~~~~~~~~~
 
-The methods in this module support Python `Buffer Protocol
-<https://docs.python.org/3/c-api/buffer.html>`__, which allows them to be used
-on any object that exports a buffer interface. Here is an example showing
-hashing of a 4D NumPy array:
+The Python `Buffer Protocol <https://docs.python.org/3/c-api/buffer.html>`__
+allows Python objects to expose their data as raw byte arrays to other objects,
+for fast access without copying to a separate location in memory.  Among
+others, NumPy is a major framework that supports this protocol.
+
+All hashing functions in this packege will read byte arrays from objects that
+expose them via the buffer protocol. Here is an example showing hashing of a 4D
+NumPy array:
 
 .. code-block:: python
 
@@ -97,8 +103,8 @@ hashing of a 4D NumPy array:
     >>> metrohash.hash64_int(arr)
     12125832280816116063
 
-Note that arrays need to be contiguous for this to work. To convert a
-non-contiguous array, use ``np.ascontiguousarray()`` method.
+The arrays need to be contiguous for this to work. To convert a non-contiguous
+array, use NumPy's ``ascontiguousarray()`` function.
 
 Development
 -----------

cpp.mk

@@ -2,7 +2,7 @@ CXX := g++
 CXXFLAGS := -std=c++11 -O3 -msse4.2
 LDFLAGS :=
 SRCEXT := cc
-INC := -I include
+INC := -I src
 LIB := -L lib
 
 INPUT := ./data/sample_100k.txt

pip-freeze.txt

@@ -8,7 +8,7 @@ ipdb==0.13.9
 ipython==7.30.1
 jedi==0.18.1
 matplotlib-inline==0.1.3
--e git+https://github.com/escherba/python-metrohash@03ec8e4b0b21bf4a9726b3625d5ebc6e791b2a82#egg=metrohash
+-e git+https://github.com/escherba/python-metrohash@cffe9bc1c0b48c2269c9e0abe1fb564ccf86d41f#egg=metrohash
 numpy==1.21.5
 packaging==21.3
 parso==0.8.3
@@ -18,7 +18,7 @@ pluggy==1.0.0
 prompt-toolkit==3.0.24
 ptyprocess==0.7.0
 py==1.11.0
-Pygments==2.11.0
+Pygments==2.11.1
 pyparsing==3.0.6
 pytest==6.2.5
 toml==0.10.2

setup.py

@@ -31,7 +31,9 @@ def is_pure(self):
 
 CXXFLAGS = []
 
-print(f"building for platform: {os.name}")
+print("building for platform: %s" % os.name)
+print("available CPU flags: %s" % CPU_FLAGS)
+
 if os.name == "nt":
     CXXFLAGS.extend(["/O2"])
 else:
@@ -42,58 +44,58 @@ def is_pure(self):
     ])
 
 
+if 'ssse3' in CPU_FLAGS:
+    print("Compiling with SSSE3 enabled")
+    CXXFLAGS.append('-mssse3')
+else:
+    print("compiling without SSE3 support")
+
+
 if 'sse4_2' in CPU_FLAGS:
     print("Compiling with SSE4.2 enabled")
     CXXFLAGS.append('-msse4.2')
 else:
     print("compiling without SSE4.2 support")
 
 
-INCLUDE_DIRS = ['include']
+INCLUDE_DIRS = ['src']
 CXXHEADERS = [
-    "include/metro.h",
-    "include/metrohash.h",
-    "include/metrohash128.h",
-    "include/metrohash128crc.h",
-    "include/metrohash64.h",
-    "include/platform.h",
+    "src/metro.h",
+    "src/metrohash.h",
+    "src/metrohash128.h",
+    "src/metrohash128crc.h",
+    "src/metrohash64.h",
+    "src/platform.h",
 ]
 CXXSOURCES = [
     "src/metrohash64.cc",
     "src/metrohash128.cc",
 ]
 
-CMDCLASS = {}
 EXT_MODULES = []
 
 if USE_CYTHON:
     print("building extension using Cython")
-    CMDCLASS['build_ext'] = build_ext
-    EXT_MODULES.append(
-        Extension(
-            "metrohash",
-            CXXSOURCES + ["src/metrohash.pyx"],
-            depends=CXXHEADERS,
-            language="c++",
-            extra_compile_args=CXXFLAGS,
-            include_dirs=INCLUDE_DIRS,
-        )
-    )
+    CMDCLASS = {'build_ext': build_ext}
+    SRC_EXT = ".pyx"
 else:
     print("building extension w/o Cython")
-    EXT_MODULES.append(
-        Extension(
-            "metrohash",
-            CXXSOURCES + ["src/metrohash.cpp"],
-            depends=CXXHEADERS,
-            language="c++",
-            extra_compile_args=CXXFLAGS,
-            include_dirs=INCLUDE_DIRS,
-        )
-    )
-
-
-VERSION = '0.1.1.post2'
+    CMDCLASS = {}
+    SRC_EXT = ".cpp"
+
+
+EXT_MODULES = [
+    Extension(
+        "metrohash",
+        CXXSOURCES + ["src/metrohash" + SRC_EXT],
+        depends=CXXHEADERS,
+        language="c++",
+        extra_compile_args=CXXFLAGS,
+        include_dirs=INCLUDE_DIRS,
+    ),
+]
+
+VERSION = '0.1.1.post3'
 URL = "https://github.com/escherba/python-metrohash"