kosuzu/cmd/txt/pull.ml

let writer accum data =
	Buffer.add_string accum data;
	String.length data

let getContent connection url =
	Curl.set_url connection url;
	Curl.perform connection

let curl_pull url =
	let result = Buffer.create 4069
	and errorBuffer = ref "" in
	let connection = Curl.init () in
	try
		Curl.set_errorbuffer connection errorBuffer;
		Curl.set_writefunction connection (writer result);
		Curl.set_followlocation connection true;
		Curl.set_url connection url;
		Curl.perform connection;
		Curl.cleanup connection;
		Ok result
	with
	| Curl.CurlException (_reason, _code, _str) ->
		Curl.cleanup connection;
		Error (Printf.sprintf "Error: %s %s" url !errorBuffer)
	| Failure s ->
		Curl.cleanup connection;
		Error (Printf.sprintf "Caught exception: %s" s)

let newer time id dir =
	match Kosuzu.File_store.to_text @@ Filename.(concat dir (Kosuzu.Id.short id) ^ ".txt") with
	| Error x -> prerr_endline x; true
	| Ok txt -> time > (Kosuzu.(Header_pack.date (Date.listing txt.date)))
	| exception (Sys_error _) -> true

let print_peers p =
	let open Kosuzu.Header_pack in
	match Msgpck.to_list p.peers with [] -> ()
	| ps -> print_endline @@
		List.fold_left (fun a x -> Printf.sprintf "%s %s" a (Msgpck.to_string x)) "peers: " ps

type filter_t = { authors: Kosuzu.Person.Set.t; topics: Kosuzu.String_set.t }

let print_pull_start width total title dir =
	Printf.printf "%*d/%s %s => %s %!" width 0 total title dir

let print_pull width total i =
	Printf.printf "\r%*d/%s %!" width (i+1) total

let printers total title dir =
	let width = String.length total in
	print_pull_start width total title dir;
	print_pull width total

let fname dir text = Filename.concat dir (Kosuzu.Text.short_id text ^ ".txt")

let pull_text url dir id =
	let u = Filename.concat url ((Kosuzu.Id.short id) ^ ".txt") in
	match curl_pull u with
	| Error msg -> Printf.eprintf "Failed getting %s: %s" u msg
	| Ok txt -> let txt = Buffer.contents txt in
		match Kosuzu.Text.of_string txt with
		| Error s -> prerr_endline s
		| Ok text ->
			let file = open_out_gen [Open_creat; Open_trunc; Open_wronly] 0o640 (fname dir text) in
			output_string file txt; close_out file

let per_text url dir filter print i id time title authors topics _refs _reps = match id with
	| "" -> Printf.eprintf "\nInvalid id for %s\n" title
	| id -> let open Kosuzu in
		print i;
		if newer time id dir
		&& (String_set.empty = filter.topics
			|| String_set.exists (fun t -> List.mem t topics) filter.topics)
		&& (Person.Set.empty = filter.authors
			|| Person.Set.exists (fun t -> List.mem (Person.to_string t) authors) filter.authors)
		then pull_text url dir id

let pull_index url authors_opt topics_opt =
	let index_url = Filename.concat url "index.pck" in
	match curl_pull index_url with
	| Error s -> prerr_endline s; false
	| Ok body ->
		match Kosuzu.Header_pack.of_string (Buffer.contents body) with
		| Error s -> Printf.printf "Error with %s: %s\n" url s; false
		| Ok pk when pk.info.id = "" ->
			Printf.printf "Empty ID index.pck, skipping %s\n" url; false
		| Ok pk when not (Kosuzu.Validate.validate_id_length pk.info.id) ->
			Printf.printf "Index pack ID longer than 32 characters, skipping %s\n" url; false
		| Ok pk when not (Kosuzu.Validate.validate_id_chars pk.info.id) ->
			Printf.printf "Index pack contains invalid ID characters, skipping %s\n" url; false
		| Ok pk ->
			let dir = Filename.concat Kosuzu.Peers.text_dir pk.info.id in
			Kosuzu.File_store.with_dir dir;
			let file = open_out_gen [Open_creat; Open_trunc; Open_wronly] 0o640
				(Filename.concat dir "index.pck") in
			output_string file ( Kosuzu.Header_pack.string {
					pk with info = { pk.info with locations = url::pk.info.locations }});
			close_out file;
			let filter = let open Kosuzu in {
				authors = (match authors_opt with Some s -> Person.Set.of_string s | None -> Person.Set.empty);
				topics =( match topics_opt with Some s -> String_set.of_string s | None -> String_set.empty);
			} in
			let name = match pk.info.title with "" -> url | title -> title in
			let print = printers (string_of_int @@ Kosuzu.Header_pack.numof_texts pk) name dir in
			try Kosuzu.Header_pack.iteri (per_text url dir filter print) pk; print_newline (); true
			with Invalid_argument msg -> Printf.printf "\nFailed to parse %s: %s\n%!" url msg; false

let pull_list auths topics =
	Curl.global_init Curl.CURLINIT_GLOBALALL;
	let pull got_one peer_url = if got_one then got_one else
		(pull_index peer_url auths topics) in
	let open Kosuzu in
	let fold_locations init peer =
		ignore @@ List.fold_left pull init peer.Peers.pack.Header_pack.info.locations;
		false
	in
	ignore @@ Peers.fold fold_locations false;
	Curl.global_cleanup ()

let pull url auths topics = match url with
	| "" -> pull_list auths topics | x -> ignore (pull_index x auths topics)

open Cmdliner
let authors = Arg.(value & opt (some string) None & info ["a"; "authors"] ~docv:"Comma-separated names" ~doc:"Filter by authors")
let topics = Arg.(value & opt (some string) None & info ["t"; "topics"] ~docv:"Comma-separated topics" ~doc:"Filter by topics")
let url = Arg.(value & pos 0 string "" & info [] ~docv:"URL" ~doc:"Repository location")

let pull_t = Term.(const pull $ url $ authors $ topics)

let cmd =
  let doc = "Pull listed texts" in
  let man = [
      `S Manpage.s_description;
      `P "Pull texts from known repositories." ]
  in
  let info = Cmd.info "pull" ~doc ~man in
  Cmd.v info pull_t