Error while training about parserator HOT 14 CLOSED

@cathydeng It worked! I already added and trained new data that I want. Thanks so much, you literally saved my day :)

from parserator.

hey @unrealwill - thanks for reporting. if you need a solution sooner while we work on fixing this, try installing parserator version 0.4.1. I think 0.4.1 may have had python 3 compatibility issues, but definitely worked w/ python 2.

@fgregg I checked & confirmed that training is currently broken for both python 2 & python 3. I checked out my last commit & it worked then, so it looks like it was introduced in one of your recent changes. thoughts as to why? also, we should prob have tests for training...😝

from parserator.

sorry, don't have time to look at this right now

On Wed, Jun 22, 2016 at 2:21 PM, Cathy Deng [email protected]
wrote:

hey @unrealwill https://github.com/unrealwill - thanks for reporting.
if you need a solution sooner while we work on fixing this, try installing
parserator version 0.4.1. I think 0.4.1 may have had python 3 compatibility
issues, but definitely worked w/ python 2.

@fgregg https://github.com/fgregg I checked & confirmed that training
is currently broken for both python 2 & python 3. I checked out my last
commit & it worked then, so it looks like it was introduced in one of your
recent changes. thoughts as to why? also, we should prob have tests for
training...😝

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#29 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAgxbVCYZbeEjJtj7Ih2x-lpbkyMuHyRks5qOYs0gaJpZM4I8Frj
.

773.888.2718

from parserator.

@cathydeng Thank you for confirming the bug, I was worrying I had made a noob mistake. I'll give 0.4.1 a try tomorrow.
@fgregg No problem I'm not in a hurry.

from parserator.

@cathydeng I also have the similar issue while doing train, I changed the version to 0.4.1 but still not working. I am using Anaconda, python 2.7.
This is the error I am getting

Traceback (most recent call last):
File "/Users/qinnanchen/anaconda/bin/parserator", line 11, in
sys.exit(dispatch())
File "/Users/qinnanchen/anaconda/lib/python2.7/site-packages/parserator/main.py", line 37, in dispatch
args.func(args)
File "/Users/qinnanchen/anaconda/lib/python2.7/site-packages/parserator/main.py", line 66, in train
training.train(module, train_file_list)
File "/Users/qinnanchen/anaconda/lib/python2.7/site-packages/parserator/training.py", line 88, in train
model_path = module.name+'/'+module.MODEL_FILE
AttributeError: 'module' object has no attribute 'MODEL_FILE'

Thank you!

from parserator.

hey @oliviac12 - what command did you run? did you include the parser name when you tried to train?

from parserator.

hey @cathydeng Thanks for the quick response. Yes I was just doing the testing code in the probablypeople page. This is the command I ran

parserator train name_data/labeled/person_labeled.xml,name_data/labeled/company_labeled.xml probablepeople

But then I read your comment earlier saying you confirmed train is broken right now, so I am in the middle of the process to go through the commit history to see if I can figure it out. If you can help that would be great!

from parserator.

you can manually pip install parserator version 0.4.1 to avoid going back in the commit history - let me know if it still doesn't work w/ the older parserator version

from parserator.

I tried parserator 0.4.1 last night actually and still getting the same error after read what you said before. But I will try again and let you know. thanks!

from parserator.

@cathydeng doesn't seem like 0.4.1 works for me.

Olivias-MacBook-Air:probablepeople qinnanchen$ pip install parserator==0.4.1
Collecting parserator==0.4.1
Requirement already satisfied (use --upgrade to upgrade): future>=0.14.3 in /Users/qinnanchen/anaconda/lib/python2.7/site-packages (from parserator==0.4.1)
Requirement already satisfied (use --upgrade to upgrade): lxml>=3.4.1 in /Users/qinnanchen/anaconda/lib/python2.7/site-packages (from parserator==0.4.1)
Requirement already satisfied (use --upgrade to upgrade): python-crfsuite>=0.7 in /Users/qinnanchen/anaconda/lib/python2.7/site-packages (from parserator==0.4.1)
Installing collected packages: parserator
Successfully installed parserator-0.4.1
Olivias-MacBook-Air:probablepeople qinnanchen$ parserator train name_data/labeled/person_labeled.xml,name_data/labeled/company_labeled.xml probablepeople
/Users/qinnanchen/probablepeople/probablepeople/init.py:64: UserWarning: You must train the model (parserator train [traindata] [modulename]) to create the generic_learned_settings.crfsuite file before you can use the parse and tag methods
warnings.warn('You must train the model (parserator train [traindata] [modulename]) to create the %s file before you can use the parse and tag methods' % MODEL_FILES[model_type])
/Users/qinnanchen/probablepeople/probablepeople/init.py:64: UserWarning: You must train the model (parserator train [traindata] [modulename]) to create the person_learned_settings.crfsuite file before you can use the parse and tag methods
warnings.warn('You must train the model (parserator train [traindata] [modulename]) to create the %s file before you can use the parse and tag methods' % MODEL_FILES[model_type])
/Users/qinnanchen/probablepeople/probablepeople/init.py:64: UserWarning: You must train the model (parserator train [traindata] [modulename]) to create the company_learned_settings.crfsuite file before you can use the parse and tag methods
warnings.warn('You must train the model (parserator train [traindata] [modulename]) to create the %s file before you can use the parse and tag methods' % MODEL_FILES[model_type])
Traceback (most recent call last):
File "/Users/qinnanchen/anaconda/bin/parserator", line 11, in
sys.exit(dispatch())
File "/Users/qinnanchen/anaconda/lib/python2.7/site-packages/parserator/main.py", line 37, in dispatch
args.func(args)
File "/Users/qinnanchen/anaconda/lib/python2.7/site-packages/parserator/main.py", line 66, in train
training.train(module, train_file_list)
File "/Users/qinnanchen/anaconda/lib/python2.7/site-packages/parserator/training.py", line 88, in train
model_path = module.name+'/'+module.MODEL_FILE
AttributeError: 'module' object has no attribute 'MODEL_FILE'

from parserator.

ok so, looks like the most recent version of probablepeople needs the most recent version of parserator (due to some recent changes where you can have multiple models per module), which has the broken training. if you clone probablepeople at commit 6cb9f7ecc6d77496c580359cb63dd38ddfab3ae9, & use parserator v 0.4.1, it should work. sorry about this! will let you know when everything is fixed.

from parserator.

can you please tell me how to clone probablepeople at a specific commit? sorry I am relatively new to git, I googled and found git reset --hard [commit number], is that the one I should use? Thank you!

from parserator.

yup that would work. and then after that, run python setup.py develop again

from parserator.

the train command is now fixed w/ parserator v0.6.0

(the probablepeople library had an additional issue, & training for that is fixed w/ v0.5.1 & parserator v0.6.0)

from parserator.