fawazahmed0 / quran-verse-detection

A Simple Program, which takes quranic verse as input and outputs the chapter & verse No

Home Page:https://fawazahmed0.github.io/quran-verse-detection/

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Quran Verse Detection

In the name of God, who has guided me to do this work

A Simple Program, which takes quranic verse as input and outputs the chapter & verse No

This is a TensorflowJS model, which can be used in browser to detect the chapter and Verse No of a given english verse. This model depends on Universal Sentence Encoder Lite Model . It will output the specific line the specific verse it corresponds to and we can use that line number to get the chapter No and verse No

Note: Line number begins from 0 for this model, for example line number 0 corresponds to chapter 1, verse 1

Live Demo: https://fawazahmed0.github.io/quran-verse-detection/

Example(codepen):

<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs"></script>
<script src="https://cdn.jsdelivr.net/npm/@tensorflow-models/universal-sentence-encoder"></script>
<script>
// Quran text to detect chapter and verse No, you can specify any verse text here to test this code
// This is chapter 31 ,verse 14
var text1 = "And We have enjoined upon man [care] for his parents. His mother carried him, [increasing her] in weakness upon weakness, and his weaning is in two years. Be grateful to Me and to your parents; to Me is the [final] destination."
// This is chapter 112 ,verse 1
var text2 = "Say, He is Allah, [who is] One"

// It will take some time to load the above scripts, the total size of this model(including the above script and model)is around 32mb
console.log("Please wait the model is loading")
// Loading quran verse detection model
var model1 = tf.loadLayersModel("https://cdn.jsdelivr.net/gh/fawazahmed0/quran-verse-detection@master/model/model.json")
// Loading universal sentence encoder model
var model2 = use.load()

async function run(){
// Assigning the models to new variables and waiting for it to load, before proceeding
var quranmodel = await model1
var usemodel = await model2

// Embedding the text into numbers, so that model can understand
var embed = await usemodel.embed([text1,text2])
// predicting
var predictions = quranmodel.predict(embed).softmax()
// Array contaning the line number of the verse
var arr = predictions.argMax(1).arraySync()
// Printing the line number specific to the verse in the console
console.log("Line Number of verse in quran of text1 and text2: ",arr)
// Printing the probability of prediction
console.log("Probability of prediction of text1 and text2: ",predictions.max(1).arraySync())

// Creating line to [chapter,verseNo] mappings

// Array containing number of verses in chapters
var chaplength = [7,286,200,176,120,165,206,75,129,109,123,111,43,52,99,128,111,110,98,135,112,78,118,64,77,227,93,88,69,60,34,30,73,54,45,83,182,88,75,85,54,53,89,59,37,35,38,29,18,45,60,49,62,55,78,96,29,22,24,13,14,11,11,18,12,12,30,52,52,44,28,28,20,56,40,31,50,40,46,42,29,19,36,25,22,17,19,26,30,20,15,21,11,8,8,19,5,8,8,11,11,8,3,9,5,4,7,3,6,3,5,4,5,6]

var mappings = []

for(i=1;i<=114;i++)
{

for(j=1;j<=chaplength[i-1];j++){
  mappings.push([i,j])
}

}

// Printing Chapter and verse Number of the text

console.log("chapter and verse No of text1 and text2:", mappings[arr[0]],mappings[arr[1]])

}

// Calling run function
run()
</script>

Output in console:

Please wait the model is loading

Line Number of verse in quran of text1 and text2: [3482, 6221]

Probability of prediction of text1 and text2: [0.9999980926513672, 0.9985342025756836]

chapter and verse No of text1 and text2: [31, 14] [112, 1]

Data:
For quran translation data, you can use quran-api

Contributions:
Thanks to Rogério Araújo for adding live demo

Share:

Please share this repo with your family & friends and Star this repo by clicking on ⭐ button above ↗️:

About

A Simple Program, which takes quranic verse as input and outputs the chapter & verse No

https://fawazahmed0.github.io/quran-verse-detection/

License:The Unlicense


Languages

Language:Python 55.7%Language:JavaScript 44.1%Language:HTML 0.2%