Is there an easy way to remove accents from characters that are contained in an attribute?
Page 1 / 1
Hi,
The string searcher or the string pair replacer are a few options to consider.
All Transformers under the Category: Strings will be worth a look:
Hi,
here is a more dynamic solution using a PythonCaller. Modify "attribute_list" (line 16, case sensitive) to include the names of the attributes you want to be checked for accents:
-----
import fmeobjects import unicodedata as ud def rmdiacritics(char): ''' Return the base character of char, by "removing" any diacritics like accents or curls and strokes and the like. ''' desc = ud.name(unicode(char)) cutoff = desc.find(' WITH ') if cutoff != -1: desc = desc :cutoff] return ud.lookup(desc) def removeAccents(feature): attribute_list = ("name", "type", "state") for attrib in feature.getAllAttributeNames(): if attrib in attribute_list: value = feature.getAttribute(attrib) if value: value = unicode(value) new_value = ''.join(ormdiacritics(char) for char in value]) feature.setAttribute(attrib, new_value)
-----
You can also download the code here, in case the forum mangles the indents.
Example values before the PythonCaller:
Attribute(encoded: utf-8): `name' has value `François' Attribute(encoded: utf-8): `state' has value `Tørst' Attribute(encoded: utf-8): `type' has value `Salé'
After the PythonCaller:
Attribute(encoded: utf-16): `name' has value `Francois' Attribute(encoded: utf-16): `state' has value `Torst' Attribute(encoded: utf-16): `type' has value `Sale' David
Hi,
here is a more dynamic solution using a PythonCaller. Modify "attribute_list" (line 16, case sensitive) to include the names of the attributes you want to be checked for accents:
import fmeobjects
import unicodedata as ud
def rmdiacritics(char):
'''
Return the base character of char, by "removing" any
diacritics like accents or curls and strokes and the like.
'''
desc = ud.name(unicode(char))
cutoff = desc.find(' WITH ')
if cutoff != -1:
desc = descu:cutoff]
return ud.lookup(desc)
def removeAccents(feature):
attribute_list = ("name", "type", "state") # Modify as needed
for attrib in feature.getAllAttributeNames():
if attrib in attribute_list:
value = feature.getAttribute(attrib)
if value:
value = unicode(value)
new_value = ''.join(rmdiacritics(char) for char in value])
feature.setAttribute(attrib, new_value)
- Attribute(encoded: utf-8): `name' has value `François'
- Attribute(encoded: utf-8): `state' has value `Tørst'
- Attribute(encoded: utf-8): `type' has value `Salé'
- Attribute(encoded: utf-16): `name' has value `Francois'
- Attribute(encoded: utf-16): `state' has value `Torst'
- Attribute(encoded: utf-16): `type' has value `Sale'
Thank David.
THis is exactly what i am looking for. However i am new with python scripts, and still stuck in FME2011 so it looks like this code will work with 2012. Any suggestions on how to implement in 2011?
Thanks!
Hi,
seems like a good opportunity to learn some Python, then :-)
Basically, replace
- "fmeobjects" with "pyfme"
- "getAttribute" with "getUnicodeString"
- "setAttribute" with "setUnicodeString"
Untested, but I think that should be enough to get it running, I hope.
Also, take a look at the pyfme API doc in <fmedir>\\\\fmeobjects\\python\\apidocs\\index.html -- in particular the methods under FMEFeature.
David
Thanks again. yes agreed i need to learn this part of FME.
One last (and probably dumb) question. What symbol is used in the caller?
Hi,
not a dumb question when you're not familiar with the PythonCaller :-)
You should use "removeAccents" for the PythonCaller.
Hint: it's almost always the function ("def ...") that takes a parameter called "feature", which represents each feature object passed into the function. E.g.
def <name of function>(feature):
David
I use StringPairReplacer like this
Source Attribute: Name
Replacement Pairs: É E È E Ë E Ê E Ô O Ç C Â A ï I î I
Result Attributer: Name2
Hi Steph,
I agree, that is the most "native" FME solution, but it assumes that you're able to populate it with all the possible variants. If something isn't caught (like a sudden "Û" in your example), it will simply pass through and result in an error further down the line, where it might not be obvious what happened.
The beauty of the PythonCaller script is that it is a lot more future-proof, although it adds quite a bit of complexity, I must admit...
David
Thanks all for the responses. Agree that the native solution is probably the mosre proper, however my issue is that i do know all the possible input scenerios i may run into. The python script worked great. With that said, i have never used the stringpairreplacer so it is good to see how this one works as well.
This function could be used in a PythonCaller:
import unicodedata def remove_accents(input_str): nkfd_form = unicodedata.normalize('NFKD', unicode(input_str)) only_ascii = nkfd_form.encode('ASCII', 'ignore') return only_asciiHi,
here is a more dynamic solution using a PythonCaller. Modify "attribute_list" (line 16, case sensitive) to include the names of the attributes you want to be checked for accents:
import fmeobjects
import unicodedata as ud
def rmdiacritics(char):
'''
Return the base character of char, by "removing" any
diacritics like accents or curls and strokes and the like.
'''
desc = ud.name(unicode(char))
cutoff = desc.find(' WITH ')
if cutoff != -1:
desc = descu:cutoff]
return ud.lookup(desc)
def removeAccents(feature):
attribute_list = ("name", "type", "state") # Modify as needed
for attrib in feature.getAllAttributeNames():
if attrib in attribute_list:
value = feature.getAttribute(attrib)
if value:
value = unicode(value)
new_value = ''.join(rmdiacritics(char) for char in value])
feature.setAttribute(attrib, new_value)
- Attribute(encoded: utf-8): `name' has value `François'
- Attribute(encoded: utf-8): `state' has value `Tørst'
- Attribute(encoded: utf-8): `type' has value `Salé'
- Attribute(encoded: utf-16): `name' has value `Francois'
- Attribute(encoded: utf-16): `state' has value `Torst'
- Attribute(encoded: utf-16): `type' has value `Sale'
Thanks David, that solution is just awesome! I used it via a Python Caller and it worked perfectly fine!!
Cheers mate!
I noticed that I regularly kept comming back to this question because of the provided code by @david_r.
I figured there probably are more people using this code so I converted it to a custom transformer:
https://hub.safe.com/transformers/stringcleaner
I noticed that I regularly kept comming back to this question because of the provided code by @david_r.
I figured there probably are more people using this code so I converted it to a custom transformer:
I noticed that I regularly kept comming back to this question because of the provided code by @david_r.
I figured there probably are more people using this code so I converted it to a custom transformer:
Here's the same code updated for Python 3.6, @jeroenstiers
import fmeobjects
import unicodedata as ud
def rmdiacritics(char):
'''
Return the base character of char, by "removing" any
diacritics like accents or curls and strokes and the like.
'''
desc = ud.name(char)
cutoff = desc.find(' WITH ')
if cutoff != -1:
desc = desc:cutoff]
return ud.lookup(desc)
def removeAccents(feature):
attribute_list = ("name", "type", "state") # Modify as needed
for attrib in feature.getAllAttributeNames():
if attrib in attribute_list:
value = feature.getAttribute(attrib)
if value:
value = str(value)
new_value = ''.join(rrmdiacritics(char) for char in value])
feature.setAttribute(attrib, new_value)
Here's the same code updated for Python 3.6, @jeroenstiers
import fmeobjects
import unicodedata as ud
def rmdiacritics(char):
'''
Return the base character of char, by "removing" any
diacritics like accents or curls and strokes and the like.
'''
desc = ud.name(char)
cutoff = desc.find(' WITH ')
if cutoff != -1:
desc = desc:cutoff]
return ud.lookup(desc)
def removeAccents(feature):
attribute_list = ("name", "type", "state") # Modify as needed
for attrib in feature.getAllAttributeNames():
if attrib in attribute_list:
value = feature.getAttribute(attrib)
if value:
value = str(value)
new_value = ''.join(rrmdiacritics(char) for char in value])
feature.setAttribute(attrib, new_value)
Use the StringPairReplacer and paste this string below into the Replacement Pairs parameter.
Create a custom transformer and use it easily in any workbench. I've created the AccentRemover just like that :D
It's only good for french though...
à a À A â a  A ç c Ç C é e É E è e È E ê e Ê E ë e Ë E î i Î I ï i Ï I ô o Ô O ù u Ù U û u Û U ü u Ü UUse the StringPairReplacer and paste this string below into the Replacement Pairs parameter.
Create a custom transformer and use it easily in any workbench. I've created the AccentRemover just like that :D
It's only good for french though...
à a À A â a  A ç c Ç C é e É E è e È E ê e Ê E ë e Ë E î i Î I ï i Ï I ô o Ô O ù u Ù U û u Û U ü u Ü UUse the StringPairReplacer and paste this string below into the Replacement Pairs parameter.
Create a custom transformer and use it easily in any workbench. I've created the AccentRemover just like that :D
It's only good for french though...
à a À A â a  A ç c Ç C é e É E è e È E ê e Ê E ë e Ë E î i Î I ï i Ï I ô o Ô O ù u Ù U û u Û U ü u Ü UHi,
here is a more dynamic solution using a PythonCaller. Modify "attribute_list" (line 16, case sensitive) to include the names of the attributes you want to be checked for accents:
-----
import fmeobjects import unicodedata as ud def rmdiacritics(char): ''' Return the base character of char, by "removing" any diacritics like accents or curls and strokes and the like. ''' desc = ud.name(unicode(char)) cutoff = desc.find(' WITH ') if cutoff != -1: desc = desc :cutoff] return ud.lookup(desc) def removeAccents(feature): attribute_list = ("name", "type", "state") for attrib in feature.getAllAttributeNames(): if attrib in attribute_list: value = feature.getAttribute(attrib) if value: value = unicode(value) new_value = ''.join(ormdiacritics(char) for char in value]) feature.setAttribute(attrib, new_value)
-----
You can also download the code here, in case the forum mangles the indents.
Example values before the PythonCaller:
Attribute(encoded: utf-8): `name' has value `François' Attribute(encoded: utf-8): `state' has value `Tørst' Attribute(encoded: utf-8): `type' has value `Salé'
After the PythonCaller:
Attribute(encoded: utf-16): `name' has value `Francois' Attribute(encoded: utf-16): `state' has value `Torst' Attribute(encoded: utf-16): `type' has value `Sale' David
HI David,
The download file no longer appears to available.
I tried the code above up I can't get it to work, and my python isn't strong enough to see where i have gone wrong. It runs but doesn't to remove the accents :(
Here is a screen shot of my python caller
HI David,
The download file no longer appears to available.
I tried the code above up I can't get it to work, and my python isn't strong enough to see where i have gone wrong. It runs but doesn't to remove the accents :(
Here is a screen shot of my python caller
Try using the StringDiacriticRemover from the FME Hub in stead: https://hub.safe.com/publishers/safe-lab/transformers/stringdiacriticremover
HI David,
The download file no longer appears to available.
I tried the code above up I can't get it to work, and my python isn't strong enough to see where i have gone wrong. It runs but doesn't to remove the accents :(
Here is a screen shot of my python caller
Thanks will do :)