Package

.gitignore

@@ -21,6 +21,7 @@ var/
 *.egg-info/
 .installed.cfg
 *.egg
+*.eggs
 
 # Installer logs
 pip-log.txt

.travis.yml

@@ -0,0 +1,26 @@
+language: python
+addons:
+  apt:
+    sources:
+    - sourceline: "ppa:mc3man/trusty-media"
+    update: true
+    packages:
+    - ffmpeg
+    - pandoc
+
+before_install:
+    - pip install --upgrade pip
+    - pip install --upgrade codecov pypandoc pytest pytest-cov pytest-mock
+install: "pip install --no-cache-dir -e ."
+# Lets do python tests later.
+#script: py.test -vv -s --cov=$TRAVIS_BUILD_DIR/au/ --cov-report term-missing
+script: make Makefile test
+#after_success: codecov
+#deploy:
+#  provider: pypi
+#  user: username
+#  password:
+#    secure: token
+#  on:
+#    tags: true
+#    branch: master

Makefile

@@ -9,25 +9,25 @@
 # 2014-09-20 Dan Ellis dpwe@ee.columbia.edu
 
 #AUDFPRINT=python audfprint.py --skip-existing --continue-on-error
-AUDFPRINT=python audfprint.py --density 100 --skip-existing
+AUDFPRINT=audfprint --density 100 --skip-existing
 
 test: test_onecore test_onecore_precomp test_onecore_newmerge test_onecore_precomppk test_mucore test_mucore_precomp test_remove
 	rm -rf precompdir precompdir_mu
 	rm -f fpdbase*.pklz
 
 test_onecore: fpdbase.pklz
-	${AUDFPRINT} match --dbase fpdbase.pklz query.mp3
+	${AUDFPRINT} match --dbase fpdbase.pklz tests/data/query.mp3
 
 test_remove: fpdbase.pklz
-	${AUDFPRINT} remove --dbase fpdbase.pklz Nine_Lives/05-Full_Circle.mp3 Nine_Lives/01-Nine_Lives.mp3
+	${AUDFPRINT} remove --dbase fpdbase.pklz tests/data/Nine_Lives/05-Full_Circle.mp3 tests/data/Nine_Lives/01-Nine_Lives.mp3
 	${AUDFPRINT} list --dbase fpdbase.pklz
-	${AUDFPRINT} add --dbase fpdbase.pklz Nine_Lives/01-Nine_Lives.mp3 Nine_Lives/05-Full_Circle.mp3
+	${AUDFPRINT} add --dbase fpdbase.pklz tests/data/Nine_Lives/01-Nine_Lives.mp3 tests/data/Nine_Lives/05-Full_Circle.mp3
 	${AUDFPRINT} list --dbase fpdbase.pklz
-	${AUDFPRINT} match --dbase fpdbase.pklz query.mp3
+	${AUDFPRINT} match --dbase fpdbase.pklz tests/data/query.mp3
 
-fpdbase.pklz: audfprint.py audfprint_analyze.py audfprint_match.py hash_table.py
-	${AUDFPRINT} new --dbase fpdbase.pklz Nine_Lives/0*.mp3
-	${AUDFPRINT} add --dbase fpdbase.pklz Nine_Lives/1*.mp3
+fpdbase.pklz: audfprint audfprint/audfprint_analyze.py audfprint/audfprint_match.py audfprint/hash_table.py
+	${AUDFPRINT} new --dbase fpdbase.pklz tests/data/Nine_Lives/0*.mp3
+	${AUDFPRINT} add --dbase fpdbase.pklz tests/data/Nine_Lives/1*.mp3
 
 test_onecore_precomp: precompdir
 	${AUDFPRINT} new --dbase fpdbase0.pklz precompdir/Nine_Lives/0*
@@ -42,11 +42,11 @@ test_onecore_newmerge: precompdir
 	${AUDFPRINT} newmerge --dbase fpdbase2.pklz fpdbase0.pklz fpdbase1.pklz
 	${AUDFPRINT} match --dbase fpdbase2.pklz precompdir/query.afpt
 
-precompdir: audfprint.py audfprint_analyze.py audfprint_match.py hash_table.py
+precompdir: audfprint/audfprint.cli.py audfprint/audfprint_analyze.py audfprint/audfprint_match.py audfprint/hash_table.py
 	rm -rf precompdir
 	mkdir precompdir
-	${AUDFPRINT} precompute --precompdir precompdir Nine_Lives/*.mp3
-	${AUDFPRINT} precompute --precompdir precompdir --shifts 4 query.mp3
+	${AUDFPRINT} precompute --precompdir precompdir tests/data/Nine_Lives/*.mp3
+	${AUDFPRINT} precompute --precompdir precompdir --shifts 4 tests/data/query.mp3
 
 test_onecore_precomppk: precomppkdir
 	${AUDFPRINT} new --dbase fpdbase0.pklz precomppkdir/Nine_Lives/0*
@@ -55,31 +55,31 @@ test_onecore_precomppk: precomppkdir
 	${AUDFPRINT} match --dbase fpdbase1.pklz precomppkdir/query.afpk
 	rm -rf precomppkdir
 
-precomppkdir: audfprint.py audfprint_analyze.py audfprint_match.py hash_table.py
+precomppkdir: audfprint audfprint/audfprint_analyze.py audfprint/audfprint_match.py audfprint/hash_table.py
 	rm -rf precomppkdir
 	mkdir precomppkdir
-	${AUDFPRINT} precompute --precompute-peaks --precompdir precomppkdir Nine_Lives/*.mp3
-	${AUDFPRINT} precompute --precompute-peaks --precompdir precomppkdir --shifts 4 query.mp3
+	${AUDFPRINT} precompute --precompute-peaks --precompdir precomppkdir tests/data/Nine_Lives/*.mp3
+	${AUDFPRINT} precompute --precompute-peaks --precompdir precomppkdir --shifts 4 tests/data/query.mp3
 
 test_mucore: fpdbase_mu.pklz
-	${AUDFPRINT} match --dbase fpdbase_mu.pklz --ncores 4 query.mp3
+	${AUDFPRINT} match --dbase fpdbase_mu.pklz --ncores 4 tests/data/query.mp3
 
-fpdbase_mu.pklz: audfprint.py audfprint_analyze.py audfprint_match.py hash_table.py
-	${AUDFPRINT} new --dbase fpdbase_mu.pklz --ncores 4 Nine_Lives/0*.mp3
-	${AUDFPRINT} add --dbase fpdbase_mu.pklz --ncores 4 Nine_Lives/1*.mp3
+fpdbase_mu.pklz: audfprint audfprint/audfprint_analyze.py audfprint/audfprint_match.py audfprint/hash_table.py
+	${AUDFPRINT} new --dbase fpdbase_mu.pklz --ncores 4 tests/data/Nine_Lives/0*.mp3
+	${AUDFPRINT} add --dbase fpdbase_mu.pklz --ncores 4 tests/data/Nine_Lives/1*.mp3
 
 test_mucore_precomp: precompdir_mu
 	${AUDFPRINT} new --dbase fpdbase_mu0.pklz --ncores 4 precompdir_mu/Nine_Lives/0*
 	${AUDFPRINT} new --dbase fpdbase_mu.pklz --ncores 4 precompdir_mu/Nine_Lives/1*
 	${AUDFPRINT} merge --dbase fpdbase_mu.pklz fpdbase_mu0.pklz
 	${AUDFPRINT} match --dbase fpdbase_mu.pklz --ncores 4 precompdir_mu/query.afpt precompdir_mu/query.afpt precompdir_mu/query.afpt precompdir_mu/query.afpt precompdir_mu/query.afpt precompdir_mu/query.afpt precompdir_mu/query.afpt
 
-precompdir_mu: audfprint.py audfprint_analyze.py audfprint_match.py hash_table.py
+precompdir_mu: audfprint/cli.py audfprint/audfprint_analyze.py audfprint/audfprint_match.py audfprint/hash_table.py
 	rm -rf precompdir_mu
 	mkdir precompdir_mu
-	${AUDFPRINT} precompute --ncores 4 --precompdir precompdir_mu Nine_Lives/*.mp3
-	${AUDFPRINT} precompute --ncores 4 --precompdir precompdir_mu --shifts 4 query.mp3 query.mp3 query.mp3 query.mp3 query.mp3 query.mp3
+	${AUDFPRINT} precompute --ncores 4 --precompdir precompdir_mu tests/data/Nine_Lives/*.mp3
+	${AUDFPRINT} precompute --ncores 4 --precompdir precompdir_mu --shifts 4 tests/data/query.mp3 tests/data/query.mp3 tests/data/query.mp3 tests/data/query.mp3 tests/data/query.mp3 tests/data/query.mp3
 
 test_hash_mask: 
-	${AUDFPRINT} new --dbase fpdbase.pklz --hashbits 16 Nine_Lives/*.mp3
-	${AUDFPRINT} match --dbase fpdbase.pklz query.mp3
+	${AUDFPRINT} new --dbase fpdbase.pklz --hashbits 16 tests/data/Nine_Lives/*.mp3
+	${AUDFPRINT} match --dbase fpdbase.pklz tests/data/query.mp3

audfprint/__init__.py

@@ -0,0 +1,5 @@
+# coding=utf-8
+import logging
+
+LOG = logging.getLogger(__name__)
+LOG.addHandler(logging.NullHandler())

audfprint_analyze.py → audfprint/audfprint_analyze.py

@@ -21,10 +21,12 @@
 import glob
 import time
 
-import audio_read
+from . import LOG
+from . import audio_read
 # For utility, glob2hashtable
-import hash_table
-import stft
+from . import hash_table
+
+from . import stft
 
 # ############### Globals ############### #
 # Special extension indicating precomputed fingerprint
@@ -86,7 +88,7 @@ def landmarks2hashes(landmarks):
     landmarks = np.array(landmarks)
     # Deal with special case of empty landmarks.
     if landmarks.shape[0] == 0:
-      return np.zeros((0, 2), dtype=np.int32)
+        return np.zeros((0, 2), dtype=np.int32)
     hashes = np.zeros((landmarks.shape[0], 2), dtype=np.int32)
     hashes[:, 0] = landmarks[:, 0]
     hashes[:, 1] = (((landmarks[:, 1] & B1_MASK) << B1_SHIFT)
@@ -287,7 +289,8 @@ def find_peaks(self, d, sr):
         else:
             # The sgram is identically zero, i.e., the input signal was identically
             # zero.  Not good, but let's let it through for now.
-            print("find_peaks: Warning: input signal is identically zero.")
+            LOG.debug('find_peaks: Warning: input signal is identically zero.')
+            # print("find_peaks: Warning: input signal is identically zero.")
         # High-pass filter onset emphasis
         # [:-1,] discards top bin (nyquist) of sgram so bins fit in 8 bits
         sgram = np.array([scipy.signal.lfilter([1, -1],
@@ -329,7 +332,7 @@ def peaks2landmarks(self, pklist):
                 for peak in peaks_at[col]:
                     pairsthispeak = 0
                     for col2 in range(col + self.mindt,
-                                       min(scols, col + self.targetdt)):
+                                      min(scols, col + self.targetdt)):
                         if pairsthispeak < self.maxpairsperpeak:
                             for peak2 in peaks_at[col2]:
                                 if abs(peak2 - peak) < self.targetdf:
@@ -357,11 +360,13 @@ def wavfile2peaks(self, filename, shifts=None):
                 # [d, sr] = librosa.load(filename, sr=self.target_sr)
                 d, sr = audio_read.audio_read(filename, sr=self.target_sr, channels=1)
             except Exception as e:  # audioread.NoBackendError:
-                message = "wavfile2peaks: Error reading " + filename
+                message = "wavfile2peaks: Error reading %s" % filename
                 if self.fail_on_error:
-                    print(e)
+                    # print(e)
+                    LOG.error('%s', message)
                     raise IOError(message)
-                print(message, "skipping")
+                LOG.debug('%s skipping', message)
+                # print(message, "skipping")
                 d = []
                 sr = self.target_sr
             # Store duration in a global because it's hard to handle
@@ -569,13 +574,16 @@ def glob2hashtable(pattern, density=20.0):
     totdur = 0.0
     tothashes = 0
     for ix, file_ in enumerate(filelist):
-        print(time.ctime(), "ingesting #", ix, ":", file_, "...")
+        LOG.debug('Ingesting #%s: %s ...', ix, file_)
+        # print(time.ctime(), "ingesting #", ix, ":", file_, "...")
         dur, nhash = g2h_analyzer.ingest(ht, file_)
         totdur += dur
         tothashes += nhash
     elapsedtime = time.clock() - initticks
-    print("Added", tothashes, "(", tothashes / totdur, "hashes/sec) at ",
-          elapsedtime / totdur, "x RT")
+    LOG.debug('Added %s %s hashes/sec at %s x RT',
+              tothashes, (tothashes / totdur), (elapsedtime / totdur))
+    # print("Added", tothashes, "(", tothashes / totdur, "hashes/sec) at ",
+    #      elapsedtime / totdur, "x RT")
     return ht
 
 

audfprint_match.py → audfprint/audfprint_match.py

@@ -21,9 +21,11 @@
 except:
     pass
 
-import audfprint_analyze
-import audio_read
-import stft
+from . import LOG
+
+from . import audfprint_analyze
+from . import audio_read
+from . import stft
 
 
 def process_info():
@@ -40,7 +42,8 @@ def process_info():
 
 def log(message):
     """ log info with stats """
-    print('%s physmem=%s utime=%s %s' % (time.ctime(), process_info()))
+    LOG.debug('physmem=%s utime=%s %s' % process_info(), message)
+    # print('%s physmem=%s utime=%s %s' % (time.ctime(), process_info()))
 
 
 def encpowerof2(val):
@@ -368,16 +371,19 @@ def match_file(self, analyzer, ht, filename, number=None):
                 numberstring = "#%d" % number
             else:
                 numberstring = ""
-            print(time.ctime(), "Analyzed", numberstring, filename, "of",
-                  ('%.3f' % durd), "s "
-                                   "to", len(q_hashes), "hashes")
+            LOG.debug('Analyzed %s %s of %.3f s to %s hashes',
+                      numberstring, filename, durd, len(q_hashes))
+            # print(time.ctime(), "Analyzed", numberstring, filename, "of",
+            #      ('%.3f' % durd), "s "
+            #                       "to", len(q_hashes), "hashes")
         # Run query
         rslts = self.match_hashes(ht, q_hashes)
         # Post filtering
         if self.sort_by_time:
             rslts = rslts[(-rslts[:, 2]).argsort(), :]
         return rslts[:self.max_returns, :], durd, len(q_hashes)
 
+    # WTF should we do about this.
     def file_match_to_msgs(self, analyzer, ht, qry, number=None):
         """ Perform a match on a single input file, return list
             of message strings """
@@ -459,10 +465,11 @@ def illustrate_match(self, analyzer, ht, filename):
                  np.array([[x[1], x[2]] for x in mlms]).T,
                  '.-r')
         # Add title
-        plt.title(filename + " : Matched as " + ht.names[results[0][0]]
-                  + (" with %d of %d hashes" % (len(matchhashes),
-                                                len(q_hashes))))
+        title = '%s :Matched as %s (with %d of %d hashes' % (
+                ht.names[results[0][0]], len(matchhashes), len(q_hashes))
+
         # Display
+        plt.title(title)
         plt.show()
         # Return
         return results

audio_read.py → audfprint/audio_read.py

@@ -45,12 +45,12 @@
 HAVE_FFMPEG = True
 
 def wavread(filename):
-  """Read in audio data from a wav file.  Return d, sr."""
-  # Read in wav file.
-  samplerate, wave_data = wav.read(filename)
-  # Normalize short ints to floats in range [-1..1).
-  data = np.asfarray(wave_data) / 32768.0
-  return data, samplerate
+    """Read in audio data from a wav file.  Return d, sr."""
+    # Read in wav file.
+    samplerate, wave_data = wav.read(filename)
+    # Normalize short ints to floats in range [-1..1).
+    data = np.asfarray(wave_data) / 32768.0
+    return data, samplerate
 
 
 def audio_read(filename, sr=None, channels=None):