fixed #11

bootphon · Nov 12, 2019 · 2248d50 · 2248d50
1 parent c403ab0
commit 2248d50
Show file tree

Hide file tree

Showing 5 changed files with 30 additions and 44 deletions.
diff --git a/TODO.org b/TODO.org
@@ -81,7 +81,6 @@ Those messages are actually logged in debug level, should be smater to be warnon
 2016-10-12 16:33:58,373 - DEBUG - WARNING (apply-cmvn:Write():util/kaldi-holder-inl.h:54) Exception caught writing Table object: ERROR (apply-cmvn:Write():kaldi-matrix.cc:1229) Failed to write matrix to stream
 ** TODO Have completion setup during installation (or configuration?)
 ** TODO Have a test_commands module for testing command line interface
-** TODO Remove the wiki page on github, as we have now a readthedoc site
 * New specifications (0.4)
 #+begin_src python
   corpus = BuckeyeCorpusPreparator('./buckeye').prepare()

diff --git a/abkhazia/commands/abkhazia_main.py b/abkhazia/commands/abkhazia_main.py
@@ -185,7 +185,7 @@ def __call__(self):
 # can be usefull to don't catch any exception, so the full error
 # traceback error is printed. To do so, just comment the following
 # line
-#@CatchExceptions
+@CatchExceptions
 def main():
     """abkhazia main entry point in command line"""
     Abkhazia()

diff --git a/abkhazia/decode/_decoder_fmllr.py b/abkhazia/decode/_decoder_fmllr.py
@@ -66,43 +66,30 @@ def decode(decoder, graph_dir):
     if decoder.mkgraph_opts['reverse'].value:
         score_opts += ' --reverse true'
 
-    # decode_fmllr.sh must be run from a subdirectory of the input
-    # acoustic model directory (here decoder.am_dir). So we do: create
-    # a subdir in am_dir as a symlink to the target recipe_dir, run
-    # the script in it, and finally delete the symlink. Moreover the
-    # script make a speaker independant decoding so we use the tweak
-    # again for si.
-    #
-    # TODO This is an error to assume write permission in
-    # decoder.am_dir!! Instead we must copy (link) the required files
-    # to decoder.recipe_dir (as in _decoder_nnet)
-    try:
-        target_sa = os.path.join(decoder.recipe_dir, 'decode')
-        if not os.path.isdir(target_sa):
-            os.makedirs(target_sa)
-        tempdir_sa = os.path.join(decoder.am_dir, 'decode_fmllr')
-        os.symlink(target_sa, tempdir_sa)
+    target = os.path.join(decoder.recipe_dir, 'decode')
+    if not os.path.isdir(target):
+        os.makedirs(target)
 
-        target_si = os.path.join(decoder.recipe_dir, 'decode.si')
-        if not os.path.isdir(target_si):
-            os.makedirs(target_si)
-        tempdir_si = os.path.join(decoder.am_dir, 'decode_fmllr.si')
-        os.symlink(target_si, tempdir_si)
+    # link requested files to decoder.recipe_dir. final.mat is optional (here
+    # only if features type is LDA)
+    for linked in ('final.mdl', 'final.alimdl', 'tree', 'final.mat'):
+        src = os.path.join(decoder.am_dir, linked)
+        if os.path.exists(src):
+            os.symlink(src, os.path.join(decoder.recipe_dir, linked))
+        elif linked == 'final.mat':
+            pass
+        else:
+            raise IOError('acoustic model file not found: {}'.format(src))
 
-        decoder._run_command((
-            'steps/decode_fmllr.sh --nj {njobs} --cmd "{cmd}" '
-            '{decode_opts} {skip_scoring} --scoring-opts "{score_opts}" '
-            '{graph} {data} {decode}'.format(
-                njobs=decoder.njobs,
-                cmd=utils.config.get('kaldi', 'decode-cmd'),
-                decode_opts=decode_opts,
-                skip_scoring=_score.skip_scoring(decoder.score_opts),
-                score_opts=_score.format(
-                    decoder.score_opts, decoder.mkgraph_opts),
-                graph=graph_dir,
-                data=os.path.join(decoder.recipe_dir, 'data', decoder.name),
-                decode=tempdir_sa)))
-    finally:
-        # remove the two symlinks we created in input am_dir
-        utils.remove(tempdir_si)
-        utils.remove(tempdir_sa)
+    decoder._run_command((
+        'steps/decode_fmllr.sh --nj {njobs} --cmd "{cmd}" '
+        '{decode_opts} {skip_scoring} --scoring-opts "{score_opts}" '
+        '{graph} {data} {decode}'.format(
+            njobs=decoder.njobs,
+            cmd=utils.config.get('kaldi', 'decode-cmd'),
+            decode_opts=decode_opts,
+            skip_scoring=_score.skip_scoring(decoder.score_opts),
+            score_opts=_score.format(decoder.score_opts, decoder.mkgraph_opts),
+            graph=graph_dir,
+            data=os.path.join(decoder.recipe_dir, 'data', decoder.name),
+            decode=target)))
diff --git a/abkhazia/decode/_decoder_nnet.py b/abkhazia/decode/_decoder_nnet.py
@@ -66,7 +66,7 @@ def decode(decoder, graph_dir):
     if not os.path.isdir(target):
         os.makedirs(target)
 
-    # link requesteed files to decoder.recipe_dir. final.mdl is
+    # link requested files to decoder.recipe_dir. final.mdl is
     # mandatory (raise if not found), others are optional and Kaldi
     # will work without
     for linked in ('final.mdl', 'cmvn_opts', 'final.mat',
@@ -75,7 +75,7 @@ def decode(decoder, graph_dir):
         if os.path.exists(src):
             os.symlink(src, os.path.join(decoder.recipe_dir, linked))
         elif linked == 'final.mdl':
-            raise IOError('model file not found {}'.format(src))
+            raise IOError('acoustic model file not found: {}'.format(src))
         else:
             decoder.log.info('optional file not here %s', src)
 

diff --git a/abkhazia/decode/decode.py b/abkhazia/decode/decode.py
@@ -59,8 +59,8 @@ def __init__(self, corpus, lm_dir, feats_dir, am_dir, output_dir,
             'nnet': 'nnet'}[self.am_type]
         if decode_type and self._decoder_type != decode_type:
             raise IOError(
-                '''cannot setup decoder, acoustic model
-                and decoder type are not compatible''')
+                'Cannot setup decoder, acoustic model '
+                'and decoder type are not compatible')
         self._decoder = decoders[self._decoder_type]
 
         self.mkgraph_opts = _mkgraph.options()