def main():
from Datasets import DatasetManager
embeddingPath = DatasetManager.getVisualEmbeddingsFullSizeFolderPath() + "VisualGlove-Full.txt"
embeddings = DatasetManager.getWordsAndEmbeddingsFromFile(embeddingPath)
FileProcessing.saveToFile(embeddings, 'tempfullEmbeddings')
pureEmbeddings = [embeddings[w] for w in embeddings.keys()]
print("Getting TSNE embeddings")
reducedEmbeddings = getTsneEmbeddings(pureEmbeddings)
print("Saving to file...")
FileProcessing.saveToFile(reducedEmbeddings, 'tempEmbeddings')
labels = [k for k in embeddings.keys()]
visualizeEmbeddings(reducedEmbeddings, labels)
def concatenateEmbeddingsFiles(folderPath, newFileName):
embeddedWords = {}
with open(newFileName, 'w', encoding='utf-8') as newFile:
for filePath in [
p for p in os.listdir(folderPath)
if os.path.isdir(folderPath + "/" + p) == False
]:
print(filePath)
localEmbeddings = DatasetManager.getWordsAndEmbeddingsFromFile(
folderPath + "/" + filePath, asStr=True)
for i, w in enumerate(localEmbeddings):
if (w not in embeddedWords):
embeddedWords[w] = 1
newFile.write("{} {}\n".format(
w,
_embeddingsToString(localEmbeddings[w],
strEmbeddings=True)))
print("Processed lines:", len(embeddedWords.keys()))
