From 0af9cfbaabaa85a66afc1e740854780ebb11afe9 Mon Sep 17 00:00:00 2001
From: Artie <54867745+artie-inc@users.noreply.github.com>
Date: Fri, 4 Oct 2019 15:52:40 -0700
Subject: [PATCH 1/4] Change ticks --> apostrophes

There are a lot of tick marks that occur in English Common Voice that should be apostrophes

This is part of a larger problem which involves quotation marks / double quotes
---
 src/corporacreator/preprocessors/en.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/src/corporacreator/preprocessors/en.py b/src/corporacreator/preprocessors/en.py
index 0d95ef6..98456db 100644
--- a/src/corporacreator/preprocessors/en.py
+++ b/src/corporacreator/preprocessors/en.py
@@ -8,5 +8,8 @@ def en(client_id, sentence):
     Returns:
       (str): Cleaned up sentence. Returning None or a `str` of whitespace flags the sentence as invalid.
     """
-    # TODO: Clean up en data
+    ## collapse all apostrophe-like marks
+    ## e.g. common_voice_en_18441344.mp3	‘I’m not a serpent!’ --> 'I'm not a serpent!'
+    sentence = sentence.replace("’","'") # right-ticks --> apostrophes
+    sentence = sentence.replace("‘","'") # left-ticks --> apostrophes
     return sentence

From 755c08be0bee07fc5d9db7a2c077a9fee0fce441 Mon Sep 17 00:00:00 2001
From: Artie <54867745+artie-inc@users.noreply.github.com>
Date: Fri, 4 Oct 2019 16:11:41 -0700
Subject: [PATCH 2/4] Replace em-dash with dash

This is also an issue WRT hyphens... should hyphens and dashes be collapsed?
---
 src/corporacreator/preprocessors/en.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/src/corporacreator/preprocessors/en.py b/src/corporacreator/preprocessors/en.py
index 98456db..6d9c945 100644
--- a/src/corporacreator/preprocessors/en.py
+++ b/src/corporacreator/preprocessors/en.py
@@ -12,4 +12,7 @@ def en(client_id, sentence):
     ## e.g. common_voice_en_18441344.mp3	‘I’m not a serpent!’ --> 'I'm not a serpent!'
     sentence = sentence.replace("’","'") # right-ticks --> apostrophes
     sentence = sentence.replace("‘","'") # left-ticks --> apostrophes
+    ## Change em-dash to dash
+    ## e.g. common_voice_en_18607891.mp3  Nelly, come here — is it morning? --> Nelly, come here – is it morning?
+    sentence = sentence.replace("—","–")
     return sentence

From de131e4da07b0fd1af417671e31188ca2c9cdc24 Mon Sep 17 00:00:00 2001
From: Artie <54867745+artie-inc@users.noreply.github.com>
Date: Fri, 4 Oct 2019 16:41:59 -0700
Subject: [PATCH 3/4] C++ --> C plus plus

there are 160 utterances with C++ spoken in common voice english, and this user is in the `test.csv` file after `test.tsv` is passed through `import_cv2.py`, and I verified that it is spoken this way
---
 src/corporacreator/preprocessors/en.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/src/corporacreator/preprocessors/en.py b/src/corporacreator/preprocessors/en.py
index 6d9c945..2da5341 100644
--- a/src/corporacreator/preprocessors/en.py
+++ b/src/corporacreator/preprocessors/en.py
@@ -8,6 +8,8 @@ def en(client_id, sentence):
     Returns:
       (str): Cleaned up sentence. Returning None or a `str` of whitespace flags the sentence as invalid.
     """
+    if client_id == "3c8f07827758e9ce8070ed287139d6d3e6457c1c16dcb972edac76c4d4333dc9e9c428711237e2b34fa29f4d249287fd238ac884534e33075958233643bbd0a1":
+        sentence = sentence.replace("C++", "C plus plus")
     ## collapse all apostrophe-like marks
     ## e.g. common_voice_en_18441344.mp3	‘I’m not a serpent!’ --> 'I'm not a serpent!'
     sentence = sentence.replace("’","'") # right-ticks --> apostrophes

From ce93881ef3d63c591e6d7b106689e5ea7d72d3b3 Mon Sep 17 00:00:00 2001
From: Artie <54867745+artie-inc@users.noreply.github.com>
Date: Fri, 4 Oct 2019 18:14:55 -0700
Subject: [PATCH 4/4] Replace "=" to "equal to" for one utterance

I listened to this myself
---
 src/corporacreator/preprocessors/en.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/src/corporacreator/preprocessors/en.py b/src/corporacreator/preprocessors/en.py
index 2da5341..6fbed01 100644
--- a/src/corporacreator/preprocessors/en.py
+++ b/src/corporacreator/preprocessors/en.py
@@ -8,6 +8,8 @@ def en(client_id, sentence):
     Returns:
       (str): Cleaned up sentence. Returning None or a `str` of whitespace flags the sentence as invalid.
     """
+    if client_id == "f2909ec9143e4ff4792b0ccb2e109da938c1c5bfc91641165b0fd32f6caf67ad234318233affcaa4117a36212440eae89dbf06c75e2cb9d4a19ddec3663044ac":
+        sentence = sentence.replace("=", "equal to")
     if client_id == "3c8f07827758e9ce8070ed287139d6d3e6457c1c16dcb972edac76c4d4333dc9e9c428711237e2b34fa29f4d249287fd238ac884534e33075958233643bbd0a1":
         sentence = sentence.replace("C++", "C plus plus")
     ## collapse all apostrophe-like marks