rules.csv format

Draft

Any folder can contain a rules.csv files. The .htaccess in the same folder will be generated/modified, if:

1. There is no .htaccess already
2. or the .htaccess contains a ## DO NOT EDIT comment

Thus the rules.csv files will generate rules for URLs under that folder, e.g. the file example/rules.csv defines how to handle anything under https://w3id.org/example/.

rules.csv format

rules.csv is a contains a series of URL patterns, using " as quote symbol, , as separator and # as comment.

The first row MUST be the header row, starting with at least:

"src","dest","statuscode","mediatype"

• The src column contains the path that should be matched, relative to the current directory.
  • The path MUST NOT start with / or ./
  • The string . can be used to match the folder root (e.g. /example/)
  • The empty string ("" in the CSV) can be used to match the folder root without a trailing / (e.g. /example)
    • If `` is listed, without listing ., then the folder root
  • If the path starts with ^ and ends with $, then it is interpreted as a Regular Expression. Groups specified using (.*) and similar can be used as $1 etc. in the dest column.
• The dest column contains the absolute URI to redirect to. If the path is specified as a regular expression, groups can be referenced by their positions, e.g. $1 for the first group.
• The column "statuscode" is optional (per row or per file), the default is "302" (Found). It SHOULD be one of the 3xx redirection codes.
• The column "mediatype" is optional (per row or per file), the default is "" (no content negotitation)

The required and optional columns MUST be in the above order. For consistency, if the optional mediatype column is present, then the optional statuscode column MUST also be present. All columns must be declared in the header row.

Example (/example/rules.csv) contains:

"src","dest","statuscode","mediatype"
".","http://example.com/home","301",""
"me","http://example.com/about-me.html","303","text/html"
"me","http://example.com/about-me.jsonld","303","application/ld+json"
"me","http://example.com/about-me.html","303",""
"sub/folders/allowed","http://example.com/flat.html","",""
"^blog/(.*)/","http://example.com/blog/post/$1/","307",""

As a table:

src	dest	statuscode	mediatype
.	http://example.com/home	301
me	http://example.com/about-me.html	303	text/html
me	http://example.com/about-me.jsonld	303	application/ld+json
me	http://example.com/about-me.html	303
sub/folders/allowed	http://example.com/flat.html
^blog/(.*)$	http://example.com/blog/post/$1/	307

The first entry redirects https://w3id.org/example/ to http://example.com/home using a HTTP status 301 Moved Permamently

Second row redirects https://w3id.org/example/me to http://example.com/about-me.html, this time using HTTP status 303 See Other, if the client requests Accept: text/html.

Third row redirects https://w3id.org/example/me to http://example.com/about-me.jsonld, this time using HTTP status 303 See Other, if the client requests Accept: application/ld+json.

Fourth row redirects https://w3id.org/example/me to http://example.com/about-me.html, this time using HTTP status 303 See Other, if the client didn't match the previous content types. (Otherwise the server would say 404 Not Found).

Fifth row redirects https://w3id.org/example/sub/folders/allowed to http://example.com/flat.html using the default HTTP status 302 Found.

The sixth row is a regular expression, so any hits like https://w3id.org/example/blog/hello-world would redirect to http://example.com/blog/post/hello-world/ - here with a 307 Temporary Redirect