Image Host Grabber

[beechcott]

Originally Posted by andw

an old set-up for hotflick.net which may help retrieve from the current servers...

Host Label: hotflick
URL Pattern: ^http:\/\/www\.hotflick\.net\/u/v\/
Search Pattern: "ID: img"

Before I edit my host file with this information, are there any cons or trade-offs to doing so, or should it provide only improvements?

[effCup]

Originally Posted by beechcott

Before I edit my host file with this information, are there any cons or trade-offs to doing so, or should it provide only improvements?

Does anything ever provide only improvements?

a) Keep a copy of anything you're about to replace.
b) Keep a copy of anything you're about to edit in.

Have you tested a) to demonstrate that it's not working?
Have you tested b) to demonstrate that it works/works better?

Testing can be... um, testing, due to the usual requirement for rigour.

Be aware that the next time you auto-update IHG's hostfile from their server it will likely replace most/all manual edits you've made with whatever IHG has decided to support/implement.

[beechcott]

Originally Posted by effCup

Does anything ever provide only improvements?

a) Keep a copy of anything you're about to replace.
b) Keep a copy of anything you're about to edit in.

Have you tested a) to demonstrate that it's not working?
Have you tested b) to demonstrate that it works/works better?

Testing can be... um, testing, due to the usual requirement for rigour.

Be aware that the next time you auto-update IHG's hostfile from their server it will likely replace most/all manual edits you've made with whatever IHG has decided to support/implement.

I see.

I tried downloading from hotflick, and ImageHost Grabber said that it didn't find any images that it could download. I then edited the hostfile with the information from anw's post, to no avail. It still can't find any images.

Originally Posted by effCup

This worked for me; passed all tests from posts here.

Host Label: pixxxels
URL Pattern: ^https?:\/\/(?:s\d+\.)?pixxxels\.org\/image\/.+$
Search Pattern:

function(pageData, pageUrl) {
	var iUrl = pageData.match(/src=('|")(https?:\/\/s\d+\.pixxxels\.org\/.+?\.jpg)\1/);
	return iUrl ? {imgUrl: iUrl[2] + "?dl=1", status: "OK"} : {imgUrl: null, status: "ABORT"}
}

I tested it on that thread, and all of the photos in post #5 downloaded as only thumbnails.

[berangus]

Originally Posted by effCup

No. What you've done there is change it from handling either single or double quotes, to only handling double quotes. The parens. denote a group. Inside it are options, separated by the pipe sign. You're saying: either double-quote or double-quote. It should be left as it was: either single-quote or double-quote.

Interesting because it didn't work for me as written, but then ran fine after I made that change. Dunno anyway thanks all the same for all your updates & expertise.

[effCup]

Originally Posted by beechcott

I tested it on that thread, and all of the photos in post #5 downloaded as only thumbnails.

Ahh, good spotting.

Try this instead:

Host Label: pixxxels
URL Pattern: ^https?:\/\/(?:s\d+\.)?pixxxels\.org\/image\/.+$
Search Pattern:

function(pageData, pageUrl) {
	var iUrl = pageData.match(/id=('|")main-image\1 src=('|")(https?:\/\/s\d+\.pixxxels\.org\/.+?\.jpg)\2/);
	return iUrl ? {imgUrl: iUrl[3] + "?dl=1", status: "OK"} : {imgUrl: null, status: "ABORT"}
}

[effCup]

A slight change to groups in P&H:

Host Label: pimpandhost
URL Pattern: ^https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/.+?$
Search Pattern:

function(pageData, pageUrl) {
	var iUrl = pageData.match(/img class=('|")normal\1 src=('|")(https?:\/\/.+?)(_l)?(\.(gif|jpe?g|png|GIF|JPE?G|PNG))\2/);
	return iUrl ? {imgUrl:  iUrl[3] + iUrl[5], status: "OK"} : {imgUrl: null, status: "ABORT"}
}

----

edit: but this seems a much more robust approach, as it avoids messing with image filetypes/dot-suffixes.

Host Label: pimpandhost
URL Pattern: ^https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/.+?$
Search Pattern:

function(pageData, pageUrl) {
	if (!pageUrl.match(/\?size=original/)) {
		pageUrl += "?size=original";
	}
	var req = new XMLHttpRequest();
	req.open("GET", pageUrl, false);
	req.send();
	var iUrl = (req.responseText).match(/img class=('|")original\1 src=('|")(https?:\/\/.+?)\2/);
	return iUrl ? {imgUrl:  iUrl[3], status: "OK"} : {imgUrl: null, status: "ABORT"}
}

edit2: Hmm, well it worked with some, but not all.

[effCup]

OK, this works not only with the new P&H link format but also with the old format, many of which links still remain. Some examples here.

Host Label: pimpandhost
URL Pattern: ^https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/.+?$

function(pageData, pageUrl) {
	if (pageUrl.match(/-original\.html$/)) {
		pageUrl = pageUrl.match(/https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/\d+/) + "?size=original";
	}
	else if (!pageUrl.match(/\?size=original/)) {
		pageUrl += "?size=original";
	}
	var req = new XMLHttpRequest();
	req.open("GET", pageUrl, false);
	req.send();
	var iUrl = (req.responseText).match(/img class=('|")original\1 src=('|")(https?:\/\/.+?)\2/);
	return iUrl ? {imgUrl:  iUrl[3], status: "OK"} : {imgUrl: null, status: "ABORT"}
}

[beechcott]

Originally Posted by effCup

Ahh, good spotting.

Try this instead:

Host Label: pixxxels
URL Pattern: ^https?:\/\/(?:s\d+\.)?pixxxels\.org\/image\/.+$
Search Pattern:

function(pageData, pageUrl) {
	var iUrl = pageData.match(/id=('|")main-image\1 src=('|")(https?:\/\/s\d+\.pixxxels\.org\/.+?\.jpg)\2/);
	return iUrl ? {imgUrl: iUrl[3] + "?dl=1", status: "OK"} : {imgUrl: null, status: "ABORT"}
}

Thank you. That one worked for all of the images in that thread.

Originally Posted by effCup

OK, this works not only with the new P&H link format but also with the old format, many of which links still remain. Some examples here.

Host Label: pimpandhost
URL Pattern: ^https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/.+?$

function(pageData, pageUrl) {
	if (pageUrl.match(/-original.html/)) {
		pageUrl = pageUrl.match(/https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/\d+/) + "?size=original";
	}
	else if (!pageUrl.match(/\?size=original/)) {
		pageUrl += "?size=original";
	}
	var req = new XMLHttpRequest();
	req.open("GET", pageUrl, false);
	req.send();
	var iUrl = (req.responseText).match(/img class=('|")original\1 src=('|")(https?:\/\/.+?)\2/);
	return iUrl ? {imgUrl:  iUrl[3], status: "OK"} : {imgUrl: null, status: "ABORT"}
}

It's not working for me. I tested it out on the page linked here, and every single PimpandHost link gave the error message "Request aborted! Can't find image on page!"

[effCup]

Originally Posted by beechcott

It's not working for me. I tested it out on the page linked here, and every single PimpandHost link gave the error message "Request aborted! Can't find image on page!"

Good testing, thanks. Try this instead:

Host Label: pimpandhost
URL Pattern: ^https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/.+?$
Search Pattern:

function(pageData, pageUrl) {
	if (pageUrl.match(/-(original|medium)\.html$/)) {
		pageUrl = pageUrl.match(/https?:\/\/(?:image\.|www\.)?pimpandhost\.com\/image\/\d+/) + "?size=original";
	}
	else if (!pageUrl.match(/\?size=original/)) {
		pageUrl += "?size=original";
	}
	var req = new XMLHttpRequest();
	req.open("GET", pageUrl, false);
	req.send();
	var iUrl = (req.responseText).match(/img class=('|")original\1 src=('|")(https?:\/\/.+?)\2/);
	return iUrl ? {imgUrl:  iUrl[3], status: "OK"} : {imgUrl: null, status: "ABORT"}
}

----
The old type of P&H link could be assigned to various views. So far we've got/have found "original" & "medium". I wonder if there were any others, maybe "small" or similar?

Please let me know if you find any further such.

[effCup]

Originally Posted by berangus

Interesting because it didn't work for me as written, but then ran fine after I made that change. Dunno anyway thanks all the same for all your updates & expertise.

Would you be kind enough to try it again, using this formulation, as written? If you still get images not downloading please link here some post(s) where that's the case, so I can test.

I don't know why you got the failure. One possibility, but I'm not saying it was the cause, may be that the host was temporarily non-responsive when you tested. That is: the "failure" may have simply looked like such. /shrug/ But I'm open to it being due to something else. Tests here have passed, using the optional single- or double-quote formulation I wrote.