Compare commits

...

1 Commits

Author SHA1 Message Date
AWieczarek
850263b3b8 IUM_06 2024-05-06 21:14:18 +02:00
5 changed files with 98 additions and 41 deletions

View File

@ -1,18 +1,18 @@
import pandas as pd
import numpy as np
import tensorflow as tf
test_data = pd.read_csv('./beer_reviews_test.csv')
X_test = test_data[['review_aroma', 'review_appearance', 'review_palate', 'review_taste']]
y_test = test_data['review_overall']
model = tf.keras.models.load_model('beer_review_sentiment_model.h5')
tokenizer = tf.keras.preprocessing.text.Tokenizer(num_words=10000)
predictions = model.predict(X_test)
print(f'Predictions shape: {predictions.shape}')
X_test_seq = tokenizer.texts_to_sequences(X_test)
X_test_pad = tf.keras.preprocessing.sequence.pad_sequences(X_test_seq, maxlen=100)
if len(predictions.shape) > 1:
predictions = predictions[:, 0]
predictions = model.predict(X_test_pad)
np.savetxt('beer_review_sentiment_predictions.csv', predictions, delimiter=',', fmt='%.10f')
results = pd.DataFrame({'Predictions': predictions, 'Actual': y_test})
results.to_csv('beer_review_sentiment_predictions.csv', index=False)

24
IUM_06-metrics.py Normal file
View File

@ -0,0 +1,24 @@
import pandas as pd
from sklearn.metrics import accuracy_score, precision_recall_fscore_support, mean_squared_error
from math import sqrt
import sys
data = pd.read_csv('beer_review_sentiment_predictions.csv')
y_pred = data['Predictions']
y_test = data['Actual']
y_test_binary = (y_test >= 3).astype(int)
build_number = sys.argv[1]
accuracy = accuracy_score(y_test_binary, y_pred.round())
precision, recall, f1, _ = precision_recall_fscore_support(y_test_binary, y_pred.round(), average='micro')
rmse = sqrt(mean_squared_error(y_test, y_pred))
print(f'Accuracy: {accuracy}')
print(f'Micro-avg Precision: {precision}')
print(f'Micro-avg Recall: {recall}')
print(f'F1 Score: {f1}')
print(f'RMSE: {rmse}')
with open(r"beer_metrics.txt", "a") as f:
f.write(f"{accuracy},{build_number}\n")

24
IUM_06-plot.py Normal file
View File

@ -0,0 +1,24 @@
import matplotlib.pyplot as plt
def main():
accuracy = []
build_numbers = []
with open("beer_metrics.txt") as f:
for line in f:
accuracy.append(float(line.split(",")[0]))
build_numbers.append(int(line.split(",")[1]))
plt.plot(build_numbers, accuracy)
plt.xlabel("Build Number")
plt.ylabel("Accuracy")
plt.title("Accuracy of the model over time")
plt.xticks(range(min(build_numbers), max(build_numbers) + 1))
plt.show()
plt.savefig("acc.png")
if __name__ == "__main__":
main()

68
Jenkinsfile vendored
View File

@ -1,51 +1,61 @@
pipeline {
agent any
agent {
dockerfile true
}
triggers {
upstream(upstreamProjects: 's464979-training/training', threshold: hudson.model.Result.SUCCESS)
}
parameters {
string(name: 'CUTOFF', defaultValue: '10000', description: 'Liczba wierszy do obcięcia ze zbioru danych')
string(name: 'KAGGLE_USERNAME', defaultValue: '', description: 'Kaggle username')
password(name: 'KAGGLE_KEY', defaultValue: '', description: 'Kaggle API key')
buildSelector(defaultSelector: lastSuccessful(), description: 'Which build to use for copying artifacts', name: 'BUILD_SELECTOR')
gitParameter branchFilter: 'origin/(.*)', defaultValue: 'training', name: 'BRANCH', type: 'PT_BRANCH'
}
stages {
stage('Clone Repository') {
steps {
git url: "https://git.wmi.amu.edu.pl/s464979/ium_464979"
git branch: 'evaluation', url: "https://git.wmi.amu.edu.pl/s464979/ium_464979"
}
}
stage('Download dataset') {
stage('Copy Dataset Artifacts') {
steps {
withEnv(["KAGGLE_USERNAME=${env.KAGGLE_USERNAME}", "KAGGLE_KEY=${env.KAGGLE_KEY}"]) {
sh "kaggle datasets download -d thedevastator/1-5-million-beer-reviews-from-beer-advocate --unzip"
}
}
}
stage('Process and Split Dataset') {
agent {
dockerfile {
filename 'Dockerfile'
reuseNode true
copyArtifacts filter: 'beer_reviews.csv,beer_reviews_train.csv,beer_reviews_test.csv', projectName: 'z-s464979-create-dataset', selector: buildParameter('BUILD_SELECTOR')
}
}
stage('Copy Training Artifacts') {
steps {
sh "chmod +x ./IUM_05-split.py"
sh "python3 ./IUM_05-split.py"
archiveArtifacts artifacts: 'beer_reviews.csv,beer_reviews_train.csv,beer_reviews_test.csv', onlyIfSuccessful: true
}
}
stage("Run") {
agent {
dockerfile {
filename 'Dockerfile'
reuseNode true
copyArtifacts filter: 'beer_review_sentiment_model.h5', projectName: 's464979-training/' + params.BRANCH, selector: buildParameter('BUILD_SELECTOR')
}
}
stage('Copy Evaluation Artifacts') {
steps {
copyArtifacts filter: 'beer_metrics.txt', projectName: '_s464979-evaluation/evaluation', selector: buildParameter('BUILD_SELECTOR'), optional: true
}
}
stage("Run predictions") {
steps {
sh "chmod +x ./IUM_05-model.py"
sh "chmod +x ./IUM_05-predict.py"
sh "python3 ./IUM_05-model.py"
sh "python3 ./IUM_05-predict.py"
archiveArtifacts artifacts: 'beer_review_sentiment_model.h5,beer_review_sentiment_predictions.csv', onlyIfSuccessful: true
archiveArtifacts artifacts: 'beer_review_sentiment_predictions.csv', onlyIfSuccessful: true
}
}
stage('Run metrics') {
steps {
sh 'chmod +x ./IUM_06-metrics.py'
sh "python3 ./IUM_06-metrics.py ${currentBuild.number}"
}
}
stage('Run plot') {
steps {
sh 'chmod +x ./IUM_06-plot.py'
sh 'python3 ./IUM_06-plot.py'
}
}
stage('Archive Artifacts') {
steps {
archiveArtifacts artifacts: '*', onlyIfSuccessful: true
}
}
}

View File

@ -1 +0,0 @@
asdasd