feat: substitute sample table selection with direct input, update TestDataset tests

aip_trainer/lambdas/lambdaGetSample.py

@@ -60,20 +60,18 @@ def lambda_handler(event, context):
         category = int(body['category'])
     except KeyError:
         category = 0
-
     language = body['language']
     try:
-        sample_idx = int(body['idx'])
+        current_transcript = str(body["transcript"])
     except KeyError:
-        sample_idx = None
-
-    app_logger.info(f"category={category}, language={language}, sample_idx={sample_idx}.")
-    lambda_df_lang = lambda_database[language]
-    current_transcript = lambda_df_lang[sample_idx] if sample_idx is not None else lambda_df_lang.get_random_sample_from_df(language, category)
+        lambda_df_lang = lambda_database[language]
+        current_transcript = lambda_df_lang.get_random_sample_from_df(language, category)
+    app_logger.info(f"category={category}, language={language}, current_transcript={current_transcript}.")
     # sentence_category = getSentenceCategory(current_transcript[0])
-    current_ipa = lambda_ipa_converter[language].convertToPhonem(current_transcript[0])
+    current_transcript = current_transcript if isinstance(current_transcript, str) else current_transcript[0]
+    current_ipa = lambda_ipa_converter[language].convertToPhonem(current_transcript)
 
-    app_logger.info(f"real_transcript={current_transcript}, ipa_transcript={current_ipa}.")
+    app_logger.info(f"real_transcript='{current_transcript}', ipa_transcript='{current_ipa}'.")
     result = {
         'real_transcript': current_transcript,
         'ipa_transcript': current_ipa,

static/css/{style-new.css → style.css}

@@ -144,7 +144,7 @@ a.disabled {
 	left: 2%;
 	top: 63%;
 	transform: translate(-0%, -0%);
-	height: 30%;
+	height: 10%;
 	width: 96%;
 	max-width: 96%;
 	background: #ffff;
@@ -152,7 +152,7 @@ a.disabled {
 	border-radius: 20px;
 	box-shadow: 0 0 20px 8px #d0d0d0;
 	overflow: scroll;
-    max-height: 30em;
+    max-height: 15%;
 }
 
 .container-small {
@@ -240,8 +240,8 @@ a.disabled {
 
 .mic-button-div {
 	position: fixed;
-	left: 60%;
-	top: 5%
+	left: 50%;
+	top: 80%
 }
 
 /*############### Drop-down ############# */
@@ -412,22 +412,22 @@ a.disabled {
 		box-shadow: 0 0 20px 8px #d0d0d0;
 	}
 
-    .container2 {
-        display: block;
-        position: absolute;
-        left: 2%;
-        top: 63%;
-        transform: translate(-0%, -0%);
-        height: 30%;
-        width: 96%;
-        max-width: 96%;
-        background: #ffff;
-        overflow: hidden;
-        border-radius: 20px;
-        box-shadow: 0 0 20px 8px #d0d0d0;
-        overflow: scroll;
-        max-height: 30em;
-    }
+	.container2 {
+		display: block;
+		position: absolute;
+		left: 2%;
+		top: 63%;
+		transform: translate(-0%, -0%);
+		height: 10%;
+		width: 96%;
+		max-width: 96%;
+		background: #ffff;
+		overflow: hidden;
+		border-radius: 20px;
+		box-shadow: 0 0 20px 8px #d0d0d0;
+		overflow: scroll;
+		max-height: 15%;
+	}
 
 	.icon-text {
 		font-size: 0.8em !important;
@@ -445,7 +445,7 @@ a.disabled {
 		/* 80px */
 		height: 3.5em;
 		padding-top: 0.4em;
-		left: 40%;
+		left: 50%;
 		line-height: 0px;
 		border: 6px solid #fff;
 		border-radius: 50%;
@@ -460,8 +460,8 @@ a.disabled {
 
 	.mic-button-div {
 		position: fixed;
-		left: 40%;
-		top: 5%
+		left: 50%;
+		top: 80%
 	}
 
 	.link-icon-div {

static/javascript/callbacks.js

@@ -23,9 +23,6 @@ let currentSoundRecorded = false;
 let currentText, currentIpa, real_transcripts_ipa, matched_transcripts_ipa;
 let wordCategories;
 let startTime, endTime;
-let allSamples = {};
-let currentSamplesObj = {};
-var timeout = null
 
 // API related variables 
 let AILanguage = "de"; // Standard is German
@@ -180,7 +177,7 @@ const prepareUiForNextSample = async () => {
     if (soundFileBad == null)
         cacheSoundFiles();
 
-    updateScore(parseFloat(document.getElementById("pronunciation_accuracy").innerHTML));
+    updateScore(parseFloat(document.getElementById("pronunciation_accuracy").innerText));
 
     document.getElementById("main_title").innerText = "Processing new sample...";
 }
@@ -198,7 +195,6 @@ const populateSampleById = (dataById) => {
     document.getElementById("recorded_ipa_script").innerText = ""
     document.getElementById("pronunciation_accuracy").innerText = "";
     document.getElementById("single_word_ipa_pair").innerText = "Reference | Spoken"
-    // document.getElementById("section_accuracy").innerText = "| Score: " + currentScore.toString() + " - (" + currentSample.toString() + ")";
     document.getElementById("section_accuracy").innerText = `| Score: ${currentScore.toString()} - sample n: ${currentSample.toString()}`;
     currentSample += 1;
 
@@ -277,7 +273,6 @@ const changeLanguage = (language, generateNewSample = false) => {
             }
         }
     }
-    getTableFromSamples(allSamples, `${AILanguage}_sentence`);
     if (generateNewSample)
         getNextSample();
 }
@@ -326,7 +321,7 @@ const startMediaDevice = () => {
             try {
                 await fetch(apiMainPathSTS + '/GetAccuracyFromRecordedAudio', {
                     method: "post",
-                    body: JSON.stringify({ "title": currentText[0], "base64Audio": audioBase64, "language": AILanguage }),
+                    body: JSON.stringify({ "title": currentText, "base64Audio": audioBase64, "language": AILanguage }),
 
                 }).then(res => res.json()).
                     then(mediaData => {
@@ -349,7 +344,7 @@ const startMediaDevice = () => {
                         real_transcripts_ipa = mediaData.real_transcripts_ipa.split(" ")
                         matched_transcripts_ipa = mediaData.matched_transcripts_ipa.split(" ")
                         wordCategories = mediaData.pair_accuracy_category.split(" ")
-                        let currentTextWords = currentText[0].split(" ")
+                        let currentTextWords = currentText.split(" ")
 
                         coloredWords = "";
                         for (let word_idx = 0; word_idx < currentTextWords.length; word_idx++) {
@@ -406,7 +401,8 @@ const playSoundForAnswerAccuracy = async (accuracy) => {
 const playAudio = async () => {
 
     document.getElementById("main_title").innerText = "Generating sound...";
-    playWithMozillaApi(currentText[0]);
+    // console.debug(`playAudio:: currentText: `, typeof currentText, "=>", currentText, "#");
+    playWithMozillaApi(currentText);
     document.getElementById("main_title").innerText = "Current Sound was played";
 
 };
@@ -472,7 +468,8 @@ const stopRecording = () => {
 const playCurrentWord = async (word_idx) => {
 
     document.getElementById("main_title").innerText = "Generating word...";
-    playWithMozillaApi(currentText[0].split(' ')[word_idx]);
+    // console.debug(`playCurrentWord:: currentText: `, typeof currentText, "=>", currentText, "#");
+    playWithMozillaApi(currentText.split(' ')[word_idx]);
     document.getElementById("main_title").innerText = "Word was played";
 }
 
@@ -532,10 +529,6 @@ const wrapWordForIndividualPlayback = (word, word_idx) => {
 // ########## Function to initialize server ###############
 // This is to try to avoid aws lambda cold start 
 try {
-    fetch(apiMainPathSTS + '/getAllSamples').then(res => res.json()).then(dataAllSamples => {
-        populateAllSamples(dataAllSamples);
-        getTableFromSamples(dataAllSamples, `${AILanguage}_sentence`);
-    });
     fetch(apiMainPathSTS + '/GetAccuracyFromRecordedAudio', {
         method: "post",
         body: JSON.stringify({ "title": '', "base64Audio": '', "language": AILanguage }),
@@ -571,67 +564,30 @@ const initializeServer = async () => {
     }
 }
 
-const populateAllSamples = async (allData) => {
-    allSamples = allData;
-}
-
-const getTableFromSamples = async (obj, lang, isFiltered = false) => {
-    let table = document.getElementById("field-samples");
-    table.innerHTML = "";
-    let objLang = obj[lang];
-    // let lenAllSamples = Object.entries(allSamples[lang]).length;
-    // console.debug(`getTableFromSamples, isFiltered: ${isFiltered}, typeof objLang: ${typeof objLang}, lenAllSamples:${lenAllSamples}, objLang.length: `, Object.entries(objLang).length, "#", objLang);
-    for (let key2 in objLang) {
-        var tr = createTableRow(`${key2}: ${objLang[key2]} || ${lang}.`, key2, isFiltered);
-        table.appendChild(tr);
-    }
-    table.appendChild(tr);
-}
-
-const createTableRow = (contentRow, sampleIdx, isFiltered = false) => {
-    var tr = document.createElement('tr');
-    tr.append(`${contentRow}`);
-    tr.onclick = async function () {
-        await prepareUiForNextSample()
-        // console.debug(`createTableRow:: ${isFiltered}, sampleIdx: `, sampleIdx);
-        await fetch(apiMainPathSample + '/getSample', {
-            method: "post",
-            body: JSON.stringify({
-                "language": AILanguage, "idx": sampleIdx
-            }),
-        }).then(res => {
-            let res2json = res.json()
-            // console.debug(`createTableRow:: ${isFiltered}, res2json: `, typeof res2json, "=>", res2json, "#");
-            return res2json
-        }).then(dataOnRowCreation => {
-            // console.debug(`createTableRow:: ${isFiltered}, dataOnRowCreation: `, typeof dataOnRowCreation, "=>", dataOnRowCreation, "#");
-            populateSampleById(dataOnRowCreation)
-            tr.style["background-color"] = "#f0f0f0";
-        })
-    };
-    return tr;
-}
-
-const filterAllSamples = async (obj, filter, lang) => {
-    if (filter == "") {
-        currentSamplesObj = {...obj}
-    };
-    objByLAng = obj[lang];
-    const filtered = Object.entries(objByLAng).filter(([key, value]) => value.toLowerCase().includes(filter));
-    currentSamplesObj = {
-        [lang]: Object.entries(filtered).map(([key, value]) => value[1])
-    };
+const getSampleFromTextInput = async (AILanguage, textInput) => {
+    await fetch(apiMainPathSample + '/getSample', {
+        method: "post",
+        body: JSON.stringify({
+            "language": AILanguage, "transcript": textInput
+        }),
+    }).then(res => {
+        let res2json = res.json()
+        // console.debug(`getSampleFromTextInput:: res2json: `, typeof res2json, "=>", res2json, "#");
+        return res2json
+    }).then(dataOnInput => {
+        console.log(`getSampleFromTextInput:: dataOnInput: `, typeof dataOnInput, "=>", dataOnInput, "#");
+        populateSampleById(dataOnInput)
+    })
 }
 
-// todo: fix the request from the rows filtered not working
 $(document).ready(function(){
-    $("#field-filter-samples").on("keyup", function(e) {
+    $("#field-filter-samples").on("keyup", async function(e) {
         e.preventDefault();
         var keycode = (e.keyCode ? e.keyCode : e.which);
         if (keycode === 13 || e.key === 'Enter') {
-            var valueFilter = $(this).val().toLowerCase();
-            filterAllSamples(allSamples, valueFilter, `${AILanguage}_sentence`)
-            getTableFromSamples(currentSamplesObj, `${AILanguage}_sentence`, true);
+            var valueFilter = $(this).val()
+            // console.debug(`input:: valueFilter: `, typeof valueFilter, "=>", valueFilter, ", AILanguage: ", AILanguage, "#");
+            await getSampleFromTextInput(AILanguage, valueFilter);
         }
     });
 });

static/main.html

@@ -18,7 +18,7 @@
     ></script>
 
     <script src="static/javascript/callbacks.js"></script>
-    <link rel="stylesheet" href="static/css/style-new.css" />
+    <link rel="stylesheet" href="static/css/style.css" />
     <link href="https://fonts.googleapis.com/icon?family=Material+Icons" rel="stylesheet" />
 </head>
 
@@ -127,7 +127,6 @@
         <div class="container2">
             <div id="div-field-filter-samples" style="position: absolute; width: 97%; margin: 1em;">
                 <input id="field-filter-samples" type="search" class="form-control" placeholder="Write and press enter to filter">
-                <div id="field-samples">{}</div>
             </div>
         </div>
 

tests/test_dataset.py

@@ -10,7 +10,7 @@ def helper_category(category: int, threshold_min: int, threshold_max: int, n: in
         event = {'body': json.dumps({'category': category, 'language': 'de'})}
         response = lambdaGetSample.lambda_handler(event, [])
         response_dict = json.loads(response)
-        number_of_words = len(response_dict['real_transcript'][0].split())
+        number_of_words = len(response_dict['real_transcript'].split())
         try:
             assert threshold_min < number_of_words <= threshold_max
         except AssertionError: