fix(data-computer): fixed lda parameters

fix(data-computer): fixed best_topic bugs

feat(data-computer): changed output of lda.py

feat(data-computer): added word_weight of each word in lda.py

feat(data-computer):changed parameters in lda.py

fix(data-computer):removed empty documents from the model

docs(data-computer): fixed a language error

docs(data-computer):changed comments, following François's advices

docs(data-computer): commented lda.py

docs(data-computer): removed the optimisation of the parameter nb_topic

docs(data-computer): add in comment afunction to optimize nb of topics

feat(data-computer): output now includes best_topic and his weight

feat(data-computer): lda's output is now in 'value'

docs(lda):added coherence system in comment

perf(lda): hyper parameters optimisation

refactor(lda): deleted useless dependencies

perf(lda): changed 'uniformize' to reduce vocabulary size

fix: fix error in lda.py

fix(data-computer): fixed type errors / key errors ...

+ add abstract to key errors due to Istex

fix(data-computer): now number of topics depends of len(data)

fix(data-computer): fixed lda's learning

add(lda) : add lemmatizer and updated requirements

add(v1/stopwords) : saved stopwords in a folder '/stopwords/'

add(lda.py) function remove_accents

fix(lda.py) : enable ' in stop words

change(lda.py) : changed the output to avoid a parse (easier to use)

fix(lda.py) : fixed id types error for variable doc_bow

fix(lda.py) : iterated output

bootstrap ws for lda