koht="/content/drive/MyDrive/oma/2023/ERR data failid /andmed/lahti/lemmadpikk.txt"
tekst=open(koht, encoding="utf-8").readlines()
sisend=[rida.strip().split() for rida in tekst]

mudel=Word2Vec(sisend, min_count=50)
mudel.save("/content/drive/MyDrive/kursused/2023/kvantitatiivne_digihumanitaaria/proov_votmesonad/lemmad1.model")
mudel.wv.most_similar(positive="maja")

from sklearn.cluster import KMeans
klastreid=1000
kmeans=KMeans(n_clusters=klastreid, n_init="auto")
kmeans.fit(mudel.wv.get_normed_vectors())
kmeans.labels_

hoidla=[[] for nr in range(klastreid)]
kataloog="/content/drive/MyDrive/kursused/2023/kvantitatiivne_digihumanitaaria/proov_votmesonad/vastused/"
for sona, koht in zip(mudel.wv.index_to_key, kmeans.labels_):
  hoidla[koht].append(sona)
f=open(kataloog+"jaotus"+str(klastreid)+".txt", "w", encoding="utf-8")
for nr in range(klastreid):
   print(nr, file=f)
   print(" ".join(hoidla[nr]), file=f)
   print(file=f)
f.close()


from gensim.models import Word2Vec
mudel=Word2Vec.load("https://minitorn.tlu.ee/~jaagup/oma/too/23/05/word2vec/lemmad1.model")
from scipy.cluster.hierarchy import dendrogram, linkage
Z=linkage(mudel.wv.get_normed_vectors(), 'ward')