Python interface is compiling

src/python/_rpitxmodule.c

@@ -0,0 +1,30 @@
+#include <Python.h>
+#include "../RpiTx.h"
+#include "../RpiGpio.h"
+
+
+static PyObject*
+_rpitx_broadcast(PyObject* self, PyObject* args) {
+    int address;
+    int length;
+    float frequency;
+    if (!PyArg_ParseTuple(args, "iif", &address, &length, &frequency)) {
+        return NULL;
+    }
+
+    setUpReadArray((void*)address, length);
+    pitx_run(MODE_IQ, 44000, frequency, 0.0, 0, readArray, NULL);
+    Py_RETURN_NONE;
+}
+
+
+static PyMethodDef _rpitx_methods[] = {
+    {"broadcast", _rpitx_broadcast, METH_VARARGS, "Low-level broadcasting."},
+    {NULL, NULL, 0, NULL}
+};
+
+
+PyMODINIT_FUNC
+init_rpitx(void) {
+    (void) Py_InitModule("_rpitx", _rpitx_methods);
+}

src/python/rpitx.py

@@ -0,0 +1,56 @@
+"""Python interface to rpitx."""
+
+from pydub import AudioSegment
+import StringIO
+import array
+import _rpitx
+import wave
+
+
+def play_fm(file_, frequency):
+    """Play a music file over FM."""
+
+    def _reencode(file_name):
+        """Returns a file-like object reencoded to the proper WAV format."""
+        reencoded = StringIO.StringIO()
+        # AudioSegment doesn't support context managers either
+        original = AudioSegment.from_file(file_name)
+        if original.channels > 2:
+            raise ValueError('Too many channels in sound file')
+        if original.channels == 2:
+            # TODO: Support stereo. For now, just overlay into mono.
+            left, right = original.split_to_mono()
+            original = left.overlay(right)
+
+        original.export(reencoded, format='wav', bitrate='44k')
+        return reencoded
+
+    encoded_file = None
+    if isinstance(file_, str):
+        if file_.endswith('.wav'):
+            with open(file_) as raw_file:
+                # wave.open doesn't support context managers, so we need to be
+                # careful about closing the file
+                wav_file = wave.open(raw_file, 'r')
+                num_channels = wav_file.getnchannels()
+                framerate = wav_file.getframerate()
+                sample_width = wav_file.getsampwidth()
+                wav_file.close()
+
+                if (
+                        num_channels != 1
+                        or framerate != 1
+                        or sample_width != 2
+                ):
+                    encoded_file = _reencode(file_)
+                else:
+                    encoded_file = AudioSegment.from_file(file_)
+        else:
+            encoded_file = _reencode(file_)
+    else:
+        encoded_file = _reencode(file_)
+
+    raw_array = array.array('c')
+    raw_array.fromstring(str(encoded_file))
+    array_address, length = raw_array.buffer_info()
+    _rpitx.broadcast(array_address, length, frequency)