This is an example of using Node.js to serve tiles to a IIIF viewer from a ZIP archive. To get around http/https mismatches, it is probably easiest to use relative paths. For example:
http://localhost/uv.html#?manifest=/AECHO/1875_01/AECHO_18750101/manifest.json
The manifest.json file in this case would use a service id that reflects the proxy mapping. For example, if Node is mapped as ziptit through a proxy, the service id for the first page could be:
"service": {
"@context": "http://iiif.io/api/image/2/context.json",
"@id": "/zipit/?path=AECHO_18750101/1875-01-01-0001",
"profile": "http://iiif.io/api/image/2/level2.json"
}
The BASE_URL value in index.js defines where the odw.json file that specifies offsets is located. ZIP uses a directory structure that identifies all of the files contained in an archive. The approach has been to use one odw.json file per newspaper issue. This file uses a json syntax created by OurDigitalWorld to identify byte-ranges to extract content. The values are mostly self-explanatory:
{
"@id": "AECHO_18750101",
"file_size": 45633749,
"manifest_offset": 43,
"manifest_size": 5363,
"zip_offsets": [
{
"ident": "AECHO_18750101/1875-01-01-0001",
"coll_offset": 39290204,
"coll_size": 6342892,
"dir_offset": 45605830,
"dir_size": 27266,
"ztype": "blocks"
},
{
"ident": "AECHO_18750101/1875-01-01-0001",
"coll_offset": 31117809,
"coll_size": 8172339,
"dir_offset": 39212759,
"dir_size": 77389,
"ztype": "tiles"
...
To run this example, clone the reposity and use the regular npm commands:
npm install
npm start
One modification that might be worthwhile is to add caching, which is well supported in Node.js. Note that the odw.json file maps both tiles and image collections called blocks, which are used for search results. This can be ignored for IIIF. There is not a lot of error handling in this setup but hopefully it illustrates how the network interactions can work. The ZIP archives used are created with the odwHocrBlockIiif python script.