Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

feat: option for user to pass in an array and receive and array durin… #4584

Merged
merged 3 commits into from
May 16, 2023
Merged

feat: option for user to pass in an array and receive and array durin… #4584

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
b38b4c3
feat: option for user to pass in an array and receive and array durin…
olgavrou May 11, 2023
23ec6d7
Merge branch 'master' into wasm_fixes
olgavrou May 16, 2023
0797e7f
add try catch and print error around flaky test
olgavrou May 12, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion wasm/example.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,7 +63,7 @@ vwPromise.then((vw) => {

vwLogger.logCBExampleToStream(example);

model.saveModel("my_model.vw");
model.saveModelToFile("my_model.vw");

vwLogger.endLogStream();
model.delete();
Expand Down
248 changes: 197 additions & 51 deletions wasm/test/test.js
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -219,8 +219,8 @@ describe('Call WASM VWModule', () => {

// save load and continue learning and predicting
let filePath = path.join(__dirname, "save_model.vw");
model2.saveModel(filePath);
model2.loadModel(filePath);
model2.saveModelToFile(filePath);
model2.loadModelFromFile(filePath);

// DONE save load and continue learning and predicting

Expand All @@ -242,6 +242,79 @@ describe('Call WASM VWModule', () => {

model2.delete();

// same for model 3
let model3 = new vw.CbWorkspace({ args_str: "--cb_explore_adf" });
assert.equal(model3.predictionType(), vw.Prediction.Type.ActionProbs);

model3.learn(example);
model3.learn(example2);

// save load and continue learning and predicting
let filePath3 = path.join(__dirname, "save_model3.vw");
model3.saveModelToFile(filePath3);

let modelBuffer = fs.readFileSync(filePath3);
let ptr = vw.wasmModule._malloc(modelBuffer.byteLength);
let heapBytes = new Uint8Array(vw.wasmModule.HEAPU8.buffer, ptr, modelBuffer.byteLength);
heapBytes.set(new Uint8Array(modelBuffer));

model3.loadModelFromArray(ptr, modelBuffer.byteLength);
vw.wasmModule._free(ptr);

// DONE save load and continue learning and predicting

model3.learn(example);
model3.learn(example2);

// collect information from model 3
let model3_sumLoss = model3.sumLoss();
assert(model3_sumLoss > 0);

let pred3 = model3.predict(example2);
assert.equal(pred3.length, 3);
let model3_action1 = pred3[0]['action'];
let model3_score1 = pred3[0]['score'];
let model3_action2 = pred3[1]['action'];
let model3_score2 = pred3[1]['score'];
let model3_action3 = pred3[2]['action'];
let model3_score3 = pred3[2]['score'];

model3.delete();

// same for model 4
let model4 = new vw.CbWorkspace({ args_str: "--cb_explore_adf" });
assert.equal(model4.predictionType(), vw.Prediction.Type.ActionProbs);

model4.learn(example);
model4.learn(example2);

// save load and continue learning and predicting
let modelarray = model4.getModelAsArray();
let filePath4 = path.join(__dirname, "save_model4.vw");
fs.writeFileSync(filePath4, Buffer.from(modelarray));
model4.loadModelFromFile(filePath4);

// DONE save load and continue learning and predicting

model4.learn(example);
model4.learn(example2);

// collect information from model 4
let model4_sumLoss = model4.sumLoss();
assert(model4_sumLoss > 0);

let pred4 = model4.predict(example2);
assert.equal(pred4.length, 3);
let model4_action1 = pred4[0]['action'];
let model4_score1 = pred4[0]['score'];
let model4_action2 = pred4[1]['action'];
let model4_score2 = pred4[1]['score'];
let model4_action3 = pred4[2]['action'];
let model4_score3 = pred4[2]['score'];

model4.delete();


// compare action/scores between the two models
assert.equal(model1_action1, model2_action1);
assert.equal(model1_score1, model2_score1);
Expand All @@ -250,6 +323,20 @@ describe('Call WASM VWModule', () => {
assert.equal(model1_action3, model2_action3);
assert.equal(model1_score3, model2_score3);

assert.equal(model1_action1, model3_action1);
assert.equal(model1_score1, model3_score1);
assert.equal(model1_action2, model3_action2);
assert.equal(model1_score2, model3_score2);
assert.equal(model1_action3, model3_action3);
assert.equal(model1_score3, model3_score3);

assert.equal(model1_action1, model4_action1);
assert.equal(model1_score1, model4_score1);
assert.equal(model1_action2, model4_action2);
assert.equal(model1_score2, model4_score2);
assert.equal(model1_action3, model4_action3);
assert.equal(model1_score3, model4_score3);

// cleanup the file

fs.unlink(filePath, (err) => {
Expand All @@ -259,68 +346,80 @@ describe('Call WASM VWModule', () => {
console.log('File removed successfully:', filePath);
}
});

fs.unlink(filePath3, (err) => {
if (err) {
console.error('Error removing file:', err);
} else {
console.log('File removed successfully:', filePath);
}
});

fs.unlink(filePath4, (err) => {
if (err) {
console.error('Error removing file:', err);
} else {
console.log('File removed successfully:', filePath);
}
});
});

it('Check logging to a file and learning form it', () => {
let model = new vw.CbWorkspace({ args_str: "--cb_explore_adf" });
let vwLogger = new vw.VWExampleLogger();
try {
let model = new vw.CbWorkspace({ args_str: "--cb_explore_adf" });
let vwLogger = new vw.VWExampleLogger();

assert.equal(model.predictionType(), vw.Prediction.Type.ActionProbs);
assert.equal(model.predictionType(), vw.Prediction.Type.ActionProbs);

let example = {
text_context: `shared | s_1 s_2
| a_1 b_1 c_1
| a_2 b_2 c_2
| a_3 b_3 c_3
`};
let example = {
text_context: `shared | s_1 s_2
| a_1 b_1 c_1
| a_2 b_2 c_2
| a_3 b_3 c_3
`};

let filePath = path.join(__dirname, "logfile.txt");
vwLogger.startLogStream(filePath);
let filePath = path.join(__dirname, "logfile.txt");
vwLogger.startLogStream(filePath);

vwLogger.logCBExampleToStream(example);
example.labels = [{ action: 0, cost: 1.0, probability: 0.5 }]
vwLogger.logCBExampleToStream(example);
vwLogger.logCBExampleToStream(example);
example.labels = [{ action: 0, cost: 1.0, probability: 0.5 }]
vwLogger.logCBExampleToStream(example);

example.labels = [{ action: 10, cost: 1.0, probability: 0.5 }]
assert.throws(() => vwLogger.logCBExampleToStream(example));
example.labels = [{ action: 10, cost: 1.0, probability: 0.5 }]
assert.throws(() => vwLogger.logCBExampleToStream(example));

assert(model.sumLoss() === 0);
assert(model.sumLoss() === 0);

vwLogger.endLogStream();
const fileStream = fs.createReadStream(filePath);
const rl = readline.createInterface({
input: fileStream,
crlfDelay: Infinity,
output: process.stdout,
terminal: false,
});
vwLogger.endLogStream();

let ex = { text_context: "" };
const fileStream = fs.createReadStream(filePath);
const rl = readline.createInterface({
input: fileStream,
crlfDelay: Infinity,
output: process.stdout,
terminal: false,
});

rl.on('line', (line) => {
if (line.trim() === '') {
model.learnLabelInString(ex);
ex.text_context = "";
}
else {
ex.text_context = ex.text_context + line + "\n";
}
});

rl.on('close', () => {
assert(model.sumLoss() > 0);
rl.on('line', model.addLine.bind(model));

model.delete();
rl.on('close', () => {
assert(model.sumLoss() > 0);

fs.unlink(filePath, (err) => {
if (err) {
console.error('Error removing file:', err);
} else {
console.log('File removed successfully:', filePath);
}
});
});
model.delete();

fs.unlink(filePath, (err) => {
if (err) {
console.error('Error removing file:', err);
} else {
console.log('File removed successfully:', filePath);
}
});
});
} catch (err) {
console.log(vw.getExceptionMessage(err));
throw new Error("vw test failed");
}
});


Expand All @@ -338,7 +437,6 @@ describe('Call WASM VWModule', () => {
| a_3 b_3 c_3
`};


assert.throws(() => vwLogger.logCBExampleToStream(example));

let filePath = path.join(__dirname, "logfile3.txt");
Expand Down Expand Up @@ -410,12 +508,13 @@ describe('Call WASM VWModule', () => {

// save load and continue learning and predicting
let filePath = path.join(__dirname, "save_model.vw");
model1.saveModel(filePath);
model1.saveModelToFile(filePath);
model1.delete();

// combine inconsistent args with model file should throw
let model2 = new vw.CbWorkspace({ args_str: "--cats 2 --min_value 0 --max_value 1 --bandwidth 1" });

assert.throws(() => model2.loadModel(filePath));
assert.throws(() => model2.loadModelFromFile(filePath));

model2.delete();
});
Expand Down Expand Up @@ -533,4 +632,51 @@ describe('Call WASM VWModule', () => {

});

it('Check load model with array', () => {

let modelBuffer = fs.readFileSync(path.join(__dirname, "model.vw"));
let ptr = vw.wasmModule._malloc(modelBuffer.byteLength);
let heapBytes = new Uint8Array(vw.wasmModule.HEAPU8.buffer, ptr, modelBuffer.byteLength);
heapBytes.set(new Uint8Array(modelBuffer));
let model = new vw.Workspace({ model_array: [ptr, modelBuffer.byteLength] });

vw.wasmModule._free(ptr);
assert.equal(model.predictionType(), vw.Prediction.Type.Scalar);
let example = model.parse("|f 6:6.8953723e-02 10:2.4920074e-02 24:1.9822951e-02 50:1.7663756e-02 73:1.6158640e-02 121:5.0723456e-02 157:4.8291773e-02 178:6.0458232e-02 179:2.0943997e-02 188:1.3043258e-02 193:7.4816257e-02 209:4.6250634e-02 233:1.8848978e-02 236:3.4921709e-02 239:2.6007419e-02 264:3.1431116e-02 265:4.3809656e-02 267:3.5724755e-02 276:2.2626529e-02 293:1.9127907e-02 302:2.7269145e-02 307:4.2694855e-02 312:3.1664621e-02 326:2.3426855e-02 368:3.9957818e-02 433:2.1424608e-02 494:3.0670732e-02 506:2.4791485e-02 550:2.8210135e-02 567:4.9445845e-02 617:2.7873196e-02 625:3.2085080e-02 630:3.6478668e-02 631:3.7034698e-02 670:4.1411690e-02 682:4.8788730e-02 702:2.5331981e-02 724:2.4551263e-02 783:4.6463773e-02 817:3.9437063e-02 837:9.0064272e-02 842:1.8598272e-02 848:7.6375939e-02 850:5.0411887e-02 852:7.4332051e-02 855:7.8169920e-02 884:1.1030679e-01 889:9.8633111e-02 894:3.9411522e-02 905:3.7478998e-02 914:5.6504101e-02 949:4.6126790e-02 950:4.5762073e-02 963:3.2610044e-02 979:4.8457999e-02 1000:2.9386828e-02 1045:3.4139425e-02 1059:3.3603869e-02 1061:4.0301725e-02 1066:7.4160680e-02 1071:2.6853660e-02 1073:8.7932266e-02 1081:7.7701092e-02 1117:9.1598287e-02 1123:9.3790986e-02 1131:4.6108399e-02 1132:4.9031150e-02 1162:3.4282148e-02 1170:3.8612958e-02 1177:5.4951586e-02 1178:4.6940666e-02 1188:2.5121527e-02 1189:3.2896131e-02 1191:9.6172296e-02 1211:4.2716283e-02 1237:3.5444438e-02 1240:3.1929389e-02 1247:6.4616486e-02 1311:7.5592339e-02 1342:3.2629944e-02 1366:5.0296519e-02 1416:3.9530758e-02 1417:3.8943492e-02 1470:4.3616120e-02 1494:6.2730476e-02 1511:6.8593867e-02 1556:3.9865732e-02 1577:4.5643266e-02 1821:8.7437980e-02 1862:5.4120179e-02 1888:4.2459391e-02 1910:4.3156520e-02 1967:9.7915463e-02 1972:4.1025959e-02 2008:5.0531935e-02 2018:3.8799949e-02 2088:4.0381286e-02 2128:3.9645299e-02 2168:5.0549522e-02 2175:7.5826041e-02 2183:1.7829052e-01 2234:3.3989199e-02 2270:8.2511209e-02 2281:5.4877985e-02 2316:4.4665784e-02 2322:4.4940550e-02 2477:4.1533679e-02 2533:4.8195656e-02 2588:9.1189362e-02 2701:4.3949336e-02 2861:7.2961919e-02 2932:5.8073092e-02 2992:5.0242696e-02 3077:8.1162862e-02 3110:1.1454716e-01 3170:5.4857526e-02 3263:6.3250236e-02 3325:4.0466305e-02 3491:1.7403087e-01 3690:7.2856687e-02 3858:1.7263067e-01 3973:7.3958009e-02 4037:7.3799074e-02 4492:1.2360445e-01 5166:5.2890390e-02 5483:5.4483801e-02 5484:7.0126176e-02 6086:1.6554411e-01 6171:1.6998538e-01 6858:5.3109396e-02 7157:7.4319251e-02 7502:6.5168351e-02 7626:6.2204096e-02 7904:1.4150701e-01 7905:5.5091526e-02 7909:1.1336020e-01 8224:1.9635071e-01 8376:7.8653499e-02 8568:8.0502190e-02 8665:9.8245032e-02 8954:1.2403890e-01 9111:1.1121018e-01 9636:6.1581194e-02 11789:8.3692431e-02 12145:8.1212714e-02 15171:8.1602275e-02 16066:8.7211892e-02 17940:8.6479917e-02 19892:9.3631372e-02 28774:1.0198968e-01 29080:1.0360513e-01 37508:3.2965177e-01 38026:2.7250558e-01 38027:2.7823427e-01 39870:9.9310391e-02");
let prediction = model.predict(example);
assert.equal(typeof prediction, "number");
model.finishExample(example);
example.delete();
model.delete();
});

it('Check load model with array and with file throws', () => {
let modelfile = path.join(__dirname, "model.vw");
let modelBuffer = fs.readFileSync(modelfile);
let ptr = vw.wasmModule._malloc(modelBuffer.byteLength);
let heapBytes = new Uint8Array(vw.wasmModule.HEAPU8.buffer, ptr, modelBuffer.byteLength);
heapBytes.set(new Uint8Array(modelBuffer));

assert.throws(() => new vw.Workspace({ model_array: [ptr, modelBuffer.byteLength], model_file: modelfile }));

vw.wasmModule._free(ptr);
});

it('Check get array, save to file, and reload', () => {

let modelBuffer = fs.readFileSync(path.join(__dirname, "model.vw"));
let ptr = vw.wasmModule._malloc(modelBuffer.byteLength);
let heapBytes = new Uint8Array(vw.wasmModule.HEAPU8.buffer, ptr, modelBuffer.byteLength);
heapBytes.set(new Uint8Array(modelBuffer));
let model = new vw.Workspace({ model_array: [ptr, modelBuffer.byteLength] });

vw.wasmModule._free(ptr);
assert.equal(model.predictionType(), vw.Prediction.Type.Scalar);
let example = model.parse("|f 6:6.8953723e-02 10:2.4920074e-02 24:1.9822951e-02 50:1.7663756e-02 73:1.6158640e-02 121:5.0723456e-02 157:4.8291773e-02 178:6.0458232e-02 179:2.0943997e-02 188:1.3043258e-02 193:7.4816257e-02 209:4.6250634e-02 233:1.8848978e-02 236:3.4921709e-02 239:2.6007419e-02 264:3.1431116e-02 265:4.3809656e-02 267:3.5724755e-02 276:2.2626529e-02 293:1.9127907e-02 302:2.7269145e-02 307:4.2694855e-02 312:3.1664621e-02 326:2.3426855e-02 368:3.9957818e-02 433:2.1424608e-02 494:3.0670732e-02 506:2.4791485e-02 550:2.8210135e-02 567:4.9445845e-02 617:2.7873196e-02 625:3.2085080e-02 630:3.6478668e-02 631:3.7034698e-02 670:4.1411690e-02 682:4.8788730e-02 702:2.5331981e-02 724:2.4551263e-02 783:4.6463773e-02 817:3.9437063e-02 837:9.0064272e-02 842:1.8598272e-02 848:7.6375939e-02 850:5.0411887e-02 852:7.4332051e-02 855:7.8169920e-02 884:1.1030679e-01 889:9.8633111e-02 894:3.9411522e-02 905:3.7478998e-02 914:5.6504101e-02 949:4.6126790e-02 950:4.5762073e-02 963:3.2610044e-02 979:4.8457999e-02 1000:2.9386828e-02 1045:3.4139425e-02 1059:3.3603869e-02 1061:4.0301725e-02 1066:7.4160680e-02 1071:2.6853660e-02 1073:8.7932266e-02 1081:7.7701092e-02 1117:9.1598287e-02 1123:9.3790986e-02 1131:4.6108399e-02 1132:4.9031150e-02 1162:3.4282148e-02 1170:3.8612958e-02 1177:5.4951586e-02 1178:4.6940666e-02 1188:2.5121527e-02 1189:3.2896131e-02 1191:9.6172296e-02 1211:4.2716283e-02 1237:3.5444438e-02 1240:3.1929389e-02 1247:6.4616486e-02 1311:7.5592339e-02 1342:3.2629944e-02 1366:5.0296519e-02 1416:3.9530758e-02 1417:3.8943492e-02 1470:4.3616120e-02 1494:6.2730476e-02 1511:6.8593867e-02 1556:3.9865732e-02 1577:4.5643266e-02 1821:8.7437980e-02 1862:5.4120179e-02 1888:4.2459391e-02 1910:4.3156520e-02 1967:9.7915463e-02 1972:4.1025959e-02 2008:5.0531935e-02 2018:3.8799949e-02 2088:4.0381286e-02 2128:3.9645299e-02 2168:5.0549522e-02 2175:7.5826041e-02 2183:1.7829052e-01 2234:3.3989199e-02 2270:8.2511209e-02 2281:5.4877985e-02 2316:4.4665784e-02 2322:4.4940550e-02 2477:4.1533679e-02 2533:4.8195656e-02 2588:9.1189362e-02 2701:4.3949336e-02 2861:7.2961919e-02 2932:5.8073092e-02 2992:5.0242696e-02 3077:8.1162862e-02 3110:1.1454716e-01 3170:5.4857526e-02 3263:6.3250236e-02 3325:4.0466305e-02 3491:1.7403087e-01 3690:7.2856687e-02 3858:1.7263067e-01 3973:7.3958009e-02 4037:7.3799074e-02 4492:1.2360445e-01 5166:5.2890390e-02 5483:5.4483801e-02 5484:7.0126176e-02 6086:1.6554411e-01 6171:1.6998538e-01 6858:5.3109396e-02 7157:7.4319251e-02 7502:6.5168351e-02 7626:6.2204096e-02 7904:1.4150701e-01 7905:5.5091526e-02 7909:1.1336020e-01 8224:1.9635071e-01 8376:7.8653499e-02 8568:8.0502190e-02 8665:9.8245032e-02 8954:1.2403890e-01 9111:1.1121018e-01 9636:6.1581194e-02 11789:8.3692431e-02 12145:8.1212714e-02 15171:8.1602275e-02 16066:8.7211892e-02 17940:8.6479917e-02 19892:9.3631372e-02 28774:1.0198968e-01 29080:1.0360513e-01 37508:3.2965177e-01 38026:2.7250558e-01 38027:2.7823427e-01 39870:9.9310391e-02");
let prediction = model.predict(example);
assert.equal(typeof prediction, "number");
model.finishExample(example);
example.delete();
model.delete();
});
});
Loading