py: expose highest-level enry.language()

Signed-off-by: Alexander Bezzubov <bzz@apache.org>

python/README.md

@@ -1,16 +1,15 @@
 # Python bindings for enry
 
-Python bingings thoug cFFI (API, out-of-line) for calling enr Go functions though CGo wrapper.
+Python bindings through cFFI (API, out-of-line) for calling enry Go functions exposed by CGo wrapper.
 
 ## Build
 
 ```
-$ make static
-$ python enry_build.py
+$ cd .. && make static
+$ python build_enry.py
 ```
 
-Will build static library for Cgo wrapper `libenry`, then generate and build `enry.c` 
-- a CPython extension that
+Will build a static library for Cgo wrapper `libenry`, then generate and build `enry.c` - a CPython extension that provides actual bindings.
 
 ## Run
 
@@ -21,9 +20,9 @@ $ python enry.py
 ```
 
 ## TODOs
- - [ ] try ABI mode, to aviod dependency on C compiler on install (+perf test?)
- - [ ] ready `libenry.h` and generate `ffibuilder.cdef` content
- - [ ] helpers for sending/recieving Go slices to C
+ - [x] helpers for sending/receiving Go slices to C
+ - [ ] read `libenry.h` and generate `ffibuilder.cdef(...)` content
  - [ ] cover the rest of enry API
  - [ ] add `setup.py`
- - [ ] build/release automation on CI (publish on pypi)
+ - [ ] build/release automation on CI (publish on pypi)
+ - [ ] try ABI mode, to avoid dependency on C compiler on install (+perf test?)

python/build_enry.py

@@ -3,10 +3,19 @@ ffibuilder = FFI()
 
 # cdef() expects a single string declaring the C types, functions and
 # globals needed to use the shared object. It must be in valid C syntax.
+# Taken from java/shared/libenry.h
 ffibuilder.cdef("""
+typedef unsigned char GoUint8;
+typedef long long GoInt64;
+typedef GoInt64 GoInt;
+
 typedef struct { const char *p; ptrdiff_t n; } _GoString_;
 typedef _GoString_ GoString;
-typedef unsigned char GoUint8;
+
+typedef struct { void *data; GoInt len; GoInt cap; } GoSlice;
+
+
+extern GoString GetLanguage(GoString p0, GoSlice p1);
 
 /* Return type for GetLanguageByExtension */
 struct GetLanguageByExtension_return {

python/enry.py

@@ -4,73 +4,85 @@ Python library calling enry Go implementation trough cFFI (API, out-of-line) and
 
 from _c_enry import ffi, lib
 
-## Helpers
+## cgo -> ffi helpers
+def py_bytes_to_go(py_bytes: bytes):
+    c_bytes = ffi.new("char[]", len(py_bytes))
+    go_slice = ffi.new("GoSlice *", [c_bytes, len(py_bytes), len(py_bytes)])
+    return (go_slice[0], c_bytes)
 
+def py_str_to_go(py_str: str):
+    str_bytes = py_str.encode()
+    c_str = ffi.new("char[]", str_bytes)
+    go_str = ffi.new("_GoString_ *", [c_str, len(str_bytes)])
+    return (go_str[0], c_str)
 
-def go_str_to_py(go_str):
+def go_str_to_py(go_str: str):
     str_len = go_str.n
     if str_len > 0:
         return ffi.unpack(go_str.p, go_str.n).decode()
     return ""
 
-
-def py_str_to_go(py_str):
-    str_bytes = py_str.encode()
-    c_str = ffi.new("char[]", str_bytes)
-    go_str = ffi.new("_GoString_ *", [c_str, len(str_bytes)])
-    return go_str[0]
-
-
-def go_bool_to_py(go_bool):
+def go_bool_to_py(go_bool: bool):
     return go_bool == 1
 
 
-## API
-
+## API, TODO(bzz): add docstrings
+def language(filename: str, content: bytes) -> str:
+    fName, c_str = py_str_to_go(filename)
+    fContent, c_bytes = py_bytes_to_go(content)
+    guess = lib.GetLanguage(fName, fContent)
+    lang = go_str_to_py(guess)
+    return lang
 
 def language_by_extension(filename: str) -> str:
-    fName = py_str_to_go(filename)
+    fName, c_str = py_str_to_go(filename)
     guess = lib.GetLanguageByExtension(fName)
     lang = go_str_to_py(guess.r0)
     return lang
 
-
 def language_by_filename(filename: str) -> str:
-    fName = py_str_to_go(filename)
+    fName, c_str = py_str_to_go(filename)
     guess = lib.GetLanguageByFilename(fName)
     lang = go_str_to_py(guess.r0)
     return lang
 
-
 def is_vendor(filename: str) -> bool:
-    fName = py_str_to_go(filename)
+    fName, c_str = py_str_to_go(filename)
     guess = lib.IsVendor(fName)
     return go_bool_to_py(guess)
 
 
 ## Tests
-
+from collections import namedtuple
 
 def main():
+    TestFile = namedtuple("TestFile", "name, content, lang")
     files = [
-        "Parse.hs", "some.cpp", "and.go", "type.h", ".bashrc", ".gitignore"
+        TestFile("Parse.hs", b"", "Haskell"), TestFile("some.cpp", b"", "C++"), 
+        TestFile("orand.go", b"", "Go"), TestFile("type.h", b"", "C"), 
+        TestFile(".bashrc", b"", "Shell"), TestFile(".gitignore", b"", "Ignore List")
     ]
 
-    print("strategy: extension")
-    for filename in files:
-        lang = language_by_extension(filename)
-        print("file: {:10s} language: '{}'".format(filename, lang))
+    print("\nstrategy: extension")
+    for f in files:
+        lang = language_by_extension(f.name)
+        print("\tfile: {:10s} language: '{}'".format(f.name, lang))
 
     print("\nstrategy: filename")
-    for filename in files:
-        lang = language_by_filename(filename)
-        print("file: {:10s} language: '{}'".format(filename, lang))
+    for f in files:
+        lang = language_by_filename(f.name)
+        print("\tfile: {:10s} language: '{}'".format(f.name, lang))
 
     print("\ncheck: is vendor?")
-    for filename in files:
-        vendor = is_vendor(filename)
-        print("file: {:10s} vendor: '{}'".format(filename, vendor))
+    for f in files:
+        vendor = is_vendor(f.name)
+        print("\tfile: {:10s} vendor: '{}'".format(f.name, vendor))
 
+    print("\nstrategy: all")
+    for f in files:
+        lang = language(f.name, f.content)
+        print("\tfile: {:10s} language: '{}'".format(f.name, lang))
+        assert lang == f.lang, "Expected '{}' but got '{}'".format(f.lang, lang)
 
 if __name__ == "__main__":
     main()