Either detected via regex or prefixed with e.g. oclc:, isbn:, etc.

e.g. https://babel.hathitrust.org/cgi/pt?id=mdp.39015006592946&view=1up&seq=3

All seem to start with http://hdl.handle.net/2027/, e.g. http://hdl.handle.net/2027/uc1.$b13952

For Google Books, using accessInfo property:

 "accessInfo": {
  "country": "US",
  "viewability": "NO_PAGES",
  "embeddable": false,
  "publicDomain": false,
  "textToSpeechPermission": "ALLOWED",
  "epub": {
   "isAvailable": false
  },
  "pdf": {
   "isAvailable": false
  },
  "webReaderLink": "http://play.google.com/books/reader?id=example&hl=&printsec=frontcover&source=gbs_api",
  "accessViewStatus": "NONE",
  "quoteSharingAllowed": false
 }

Could be as simple as URI-decoding a single hash parameter, so you could link in with a search pre-populated.

Similar to the approach used for OCLC/LCCN/ISBN, we should probably store an array of already queried identifiers for IA<->GB lookup queries so we aren't hammering things.

This would make forward/back work for searches.

Potentially related to #13.

This may make it more accurate for book-to-book comparisons with page counts from other services.

Not sure if that's a temporary problem, or what, but I love this tool and that error seems to be preventing it from working! :)

e.g. .de etc.

There's some weird stuff in the ISBN fields. We should normalize it before checking for matches.

e.g. https://archive.org/details/oxyrhynchuspappt06grenuoft/page/236

See also #23 (which handled /stream/) and #31. Should probably just make this a general solution.

This is sometimes in field 955 $q of the MARC XML (see HathiTrust Bibliographic Metadata Specifications). We'd need to switch to full calls to the Bibliographic API to get this data. This will only be present in HT items explicitly ingested from IA.

Use the Google Books search API to see if we find matching identifiers from HT/IA results.

Performed a search for this volume: https://archive.org/details/analecticmagazi01irvigoog

And got back dozens of datatables popup warning messages similar to the below that had to be confirmed before being allowed to continue loading the page. In the end, I got the results, but it was a hairy experience in both Firefox (59.0.2 (64-bit)) and Chrome (Version 66.0.3359.139 (Official Build) (64-bit)), both on a Macbook Air.

DataTables warning: table id=table - Requested unknown parameter '3' for row 68, column 3. For more information about this error, please see http://datatables.net/tn/4

e.g. http://catalog.hathitrust.org/Record/001057888

Will probably need to do one level of dereferencing, and then query the first or all resulting identifiers.

Inverse of #12. Example searches:

• https://catalog.hathitrust.org/Record/000280889 yields https://books.google.com/books?id=LIM0AAAAMAAJ
• https://books.google.com/books?id=LIM0AAAAMAAJ yields no other matches

e.g. https://play.google.com/store/books/details?id=1d5UAAAAcAAJ = https://books.google.com/books?id=1d5UAAAAcAAJ

It looks like we'd need a separate call to the Data API (likely proxied to keep the access key private) to extract: htd:numpages­: the number of pages in a volume­type resource

Restricted volumes, e.g. https://archive.org/details/womenslettersfro800bagn, should display the lock icon, like HathiTrust and Google Books do.

Might be handy to be able to export/download results, e.g. as CSV.

Instead of just /details/.

Would be nice to be able to click on an example, and have it auto-populate + run the search for that example.

Should also match a year and/or volume in order to make it into the filtered results.

See: https://developers.google.com/books/docs/v1/reference/volumes#resource

Projection
You can use the projection parameter with one of the following values to specify a predefined set of Volume fields to return:

full - Returns all Volume fields.
lite - Returns only certain fields. See field descriptions marked with double asterisks in the Volume reference to find out which fields are included.
The following example returns search results with limited volume information:

GET https://www.googleapis.com/books/v1/volumes?q=flowers&projection=lite&key=yourAPIKey

It looks like we've been getting "lite" by default, which apparently doesn't include these.

For example (via @hadro in #8): http://babel.hathitrust.org/cgi/pt?id=mdp.39015085786880 shows OCLC http://www.worldcat.org/oclc/66269195 which has ebook versions http://www.worldcat.org/title/translation-of-the-confucian-i-ching-or-the-classic-of-change/oclc/220258848/editions?referer=di&editionsView=true#%2528x0%253Abook%2Bx4%253Adigital%2529format which associates back to HT and also to GB at: https://books.google.com/books?id=HJhKAQAAMAAJ + https://books.google.com/books?id=QAEaAQAAIAAJ

e.g. https://archive.org/details/cu31924031242450/ winds up returning hits for both itself and without the trailing slash.

This may catch good matches that no existing identifier mapping does (due to poor metadata).

Fusion Tables is going away on December 3rd, 2019: https://support.google.com/fusiontables/answer/9185417

We need another service for holding/querying the match tables.

It would be cool if you could bulk query identifiers by running the JavaScript via Node on the command line, for example.

See e.g. https://archive.org/details/bub_gb_LaUnOztbkP4C

e.g. https://www.google.com/books/edition/_/xSIAAAAAYAAJ

You should be able to pivot your search to an existing search result by e.g. clicking a "search" icon next to the result. This might help catch things where we have one-directional mappings.

Starting this issue just for collecting various types of matches that should be found, but currently aren't for whatever reason. May help with diagnosing problems or figuring out enhancements that might add these matches.

This is a cool service! (although not as cool as when I giddily thought it might be aligning texts across services :) )

Having said that, there seem to be some issues when searching for one of the books in another of your issues "Women's letters from Egypt". There are a number of non-matches included in the list from things which are plausible, like "Letters from ancient Egypt" with a similar title, but different authors, page counts, etc to "Angelina ice skates" which seems very far off the mark. I didn't look at the code to see how it's searching, but something definitely needs a little tweaking.

The bottom of the display also has a long list of errors of the form:

Error in HathiTrust AJAX call.
Error in Google Books AJAX call for identifier 185512691
Error in Google Books AJAX call for identifier 9780517596197

... with another ~20 lines like the last, but with different identifiers.

Each result should have a hover or expansion UI element which can give a description of why it was included, e.g. "ISBN identifier match from [x]"