Thank you brother, but this PR might break things for others who have already implemented the api.
For example, you have used "arabicnumber": "Introduction 8" , but arabicnumber should always be a number and not string + number. I will look into this PR later

I don't think so, it would only break if you try to change hadithnumber to a string because its saved as a number.

Arabicnumber on the other hand is always saved as a String even if the string contains a number theoretically. For example here:
"hadithnumber": 1035,
"arabicnumber": "463.01"

So even if somebody retrieves the object "arabicnumber", he's already forced to obtain it as a string.

InshaaAllah!

well, they will parse that string as float or int, and if it's "Introduction 8" it will throw error.
Also "arabicnumber": "463.01" shouldn't have been a string, it should be a number(mistake done by me). I am not sure if I could change it now. (because people have already implemented this api).

I found some more wrong references. For example this one. It is also book 0 hadith 0 in info.json.
I wanted to fix the reference of this hadith too but it basically doesn't exist on sunnah.com. And it also doesn't exist on the Sahih Muslim Book from Darussalam

The text + hadithnumber of this hadith can only be found in your urd and ben translation. Maybe they used another arabic print ?

My suggestion would be to stick to "darussalam and sunnah.com" way of referencing and I will write down all ahadith which cant be found on "darussalam and sunnah.com", so we can batch delete these hadithnumbers later and order the hadithnumbers again.
But obv this is only my suggestion and I don't mind approaching it differently.

Thats how it looks like rn unedited:

{
"hadithnumber": 127,
"arabicnumber": "21.03",
"grades": [],
"reference": {
"book": 1,
"hadith": 35
}
},
{
"hadithnumber": 128,
"arabicnumber": 0,
"grades": [],
"reference": {
"book": 0,
"hadith": 0
}
},
{
"hadithnumber": 129,
"arabicnumber": "22",
"grades": [],
"reference": {
"book": 1,
"hadith": 36
}
},

I don't intend to delete any hadith ( I usually place the hadiths which does not have proper reference into section 0 )
Also incase of sahih muslim, you should compare arabicnumber with Reference given in sunnah.com.
The hadithnumber in sahih muslim is a sequence wise reference number (used by urdu etc translations) and this referencing is still used in many websites.

Parsing issues should be solved. I replaced "Introduction 1/Introduction 2...." with "0.1/0.2...."

One issue remaining:
Hadithnumber 0 can be found inside the urdu json file but it isn't present inside the arabic json file.
It should also be present in all non urdu json files like this after adding it to info.json file:

{
			"hadithnumber": 0,
			"text": "",
			"grades": [],
			"reference": {
				"book": 0,
				"hadith": 0
			}
		},

Thank you brother, I will have to think if this will introduce any breaking changes. You need to wait a bit.
I will fix all the important issues (included this one) by end of ramadan inshallah.

well, they will parse that string as float or int, and if it's "Introduction 8" it will throw error. Also "arabicnumber": "463.01" shouldn't have been a string, it should be a number(mistake done by me). I am not sure if I could change it now. (because people have already implemented this api).

Why don't you fix all these major(code-breaking) bugs/issues in a new/major version-2?
That way, people who have already implemented this API will be able to continue using it without breaking their code and they can also move to the newer version if they like.

Ref.: Semantic versioning