url percent encoding & (naïve) fuzzy matching

gleam.toml

@@ -0,0 +1,19 @@
+name = "bahnhofname"
+version = "0.1.0"
+
+# Fill out these fields if you intend to generate HTML documentation or publish
+# your project to the Hex package manager.
+#
+# licences = ["Apache-2.0"]
+# description = "A Gleam library..."
+# repository = { type = "github", user = "username", repo = "project" }
+# links = [{ title = "Website", href = "https://gleam.run" }]
+
+[dependencies]
+gleam_stdlib = "~> 0.19"
+gleam_http = "~> 3.0"
+mist = "~> 0.4"
+gleam_hackney = "~> 1.0"
+
+[dev-dependencies]
+gleeunit = "~> 0.6"

manifest.toml

@@ -0,0 +1,28 @@
+# This file was generated by Gleam
+# You typically do not need to edit this file
+
+packages = [
+  { name = "certifi", version = "2.9.0", build_tools = ["rebar3"], requirements = [], otp_app = "certifi", source = "hex", outer_checksum = "266DA46BDB06D6C6D35FDE799BCB28D36D985D424AD7C08B5BB48F5B5CDD4641" },
+  { name = "gleam_erlang", version = "0.18.1", build_tools = ["gleam"], requirements = ["gleam_stdlib"], otp_app = "gleam_erlang", source = "hex", outer_checksum = "C69F59D086AD50B80DE294FB0963550630971C9DC04E92B1F7AEEDD2C0BE226C" },
+  { name = "gleam_hackney", version = "1.0.0", build_tools = ["gleam"], requirements = ["gleam_http", "gleam_stdlib", "hackney"], otp_app = "gleam_hackney", source = "hex", outer_checksum = "B3C1E6BD138D57252F9F9E499C741E9227EE7EE9B017CA650EC8193E02F734E1" },
+  { name = "gleam_http", version = "3.2.0", build_tools = ["gleam"], requirements = ["gleam_stdlib"], otp_app = "gleam_http", source = "hex", outer_checksum = "D034F5CE0639CD142CBA210B7D5D14236C284B0C5772A043D2E22128594573AE" },
+  { name = "gleam_otp", version = "0.5.3", build_tools = ["gleam"], requirements = ["gleam_erlang", "gleam_stdlib"], otp_app = "gleam_otp", source = "hex", outer_checksum = "6E705B69464237353E0380AC8143BDB29A3F0BF6168755D5F2D6E55A34A8B077" },
+  { name = "gleam_stdlib", version = "0.28.1", build_tools = ["gleam"], requirements = [], otp_app = "gleam_stdlib", source = "hex", outer_checksum = "73F0A89FADE5022CBEF6D6C3551F9ADCE7054AFCE0CB1DC4C6D5AB4CA62D0111" },
+  { name = "gleeunit", version = "0.10.1", build_tools = ["gleam"], requirements = ["gleam_stdlib"], otp_app = "gleeunit", source = "hex", outer_checksum = "ECEA2DE4BE6528D36AFE74F42A21CDF99966EC36D7F25DEB34D47DD0F7977BAF" },
+  { name = "glisten", version = "0.7.0", build_tools = ["gleam"], requirements = ["gleam_erlang", "gleam_stdlib", "gleam_otp"], otp_app = "glisten", source = "hex", outer_checksum = "52B530FF25370590843998D1B6C4EC6169DB1300D5E4407A5CDA1575374B7AEC" },
+  { name = "hackney", version = "1.18.1", build_tools = ["rebar3"], requirements = ["certifi", "metrics", "mimerl", "ssl_verify_fun", "idna", "parse_trans", "unicode_util_compat"], otp_app = "hackney", source = "hex", outer_checksum = "A4ECDAFF44297E9B5894AE499E9A070EA1888C84AFDD1FD9B7B2BC384950128E" },
+  { name = "idna", version = "6.1.1", build_tools = ["rebar3"], requirements = ["unicode_util_compat"], otp_app = "idna", source = "hex", outer_checksum = "92376EB7894412ED19AC475E4A86F7B413C1B9FBB5BD16DCCD57934157944CEA" },
+  { name = "metrics", version = "1.0.1", build_tools = ["rebar3"], requirements = [], otp_app = "metrics", source = "hex", outer_checksum = "69B09ADDDC4F74A40716AE54D140F93BEB0FB8978D8636EADED0C31B6F099F16" },
+  { name = "mimerl", version = "1.2.0", build_tools = ["rebar3"], requirements = [], otp_app = "mimerl", source = "hex", outer_checksum = "F278585650AA581986264638EBF698F8BB19DF297F66AD91B18910DFC6E19323" },
+  { name = "mist", version = "0.10.0", build_tools = ["gleam"], requirements = ["gleam_erlang", "gleam_stdlib", "gleam_otp", "glisten", "gleam_http"], otp_app = "mist", source = "hex", outer_checksum = "5AFBABABF738BAB8720F047471051E4E9D102CA4694C120DB899FA12AD5D180B" },
+  { name = "parse_trans", version = "3.3.1", build_tools = ["rebar3"], requirements = [], otp_app = "parse_trans", source = "hex", outer_checksum = "07CD9577885F56362D414E8C4C4E6BDF10D43A8767ABB92D24CBE8B24C54888B" },
+  { name = "ssl_verify_fun", version = "1.1.6", build_tools = ["mix", "rebar3", "make"], requirements = [], otp_app = "ssl_verify_fun", source = "hex", outer_checksum = "BDB0D2471F453C88FF3908E7686F86F9BE327D065CC1EC16FA4540197EA04680" },
+  { name = "unicode_util_compat", version = "0.7.0", build_tools = ["rebar3"], requirements = [], otp_app = "unicode_util_compat", source = "hex", outer_checksum = "25EEE6D67DF61960CF6A794239566599B09E17E668D3700247BC498638152521" },
+]
+
+[requirements]
+gleam_hackney = "~> 1.0"
+gleam_http = "~> 3.0"
+gleam_stdlib = "~> 0.19"
+gleeunit = "~> 0.6"
+mist = "~> 0.4"

src/bahnhofname.gleam

@@ -4,14 +4,113 @@ import gleam/http.{Get}
 import gleam/bit_builder.{BitBuilder}
 import gleam/erlang/process
 import gleam/io
+import gleam/int
 import gleam/string
+import gleam/bit_string
 import gleam/list
 import gleam/map.{Map}
 import gleam/result.{lazy_unwrap}
 import gleam/uri
 import gleam/hackney
+import gleam/option.{None, Some}
 import mist
 
+fn do_distlist(
+  b: String,
+  distlist: List(Int),
+  grapheme: String,
+  new_distlist: List(Int),
+  last_dist: Int,
+) {
+  case #(b, distlist) {
+    #("", _) -> list.reverse(new_distlist)
+    #(_, [distlist_hd, distlist_snd, ..distlist_tl]) -> {
+      let assert Ok(b_hd) = string.first(b)
+      let b_tl = string.drop_left(b, up_to: 1)
+      let diff = case #(b_hd, grapheme) {
+        #(a, b) if a != b -> 1
+        _ -> 0
+      }
+      let minimum =
+        int.min(int.min(last_dist + 1, distlist_snd + 1), distlist_hd + diff)
+      do_distlist(
+        b_tl,
+        [distlist_snd, ..distlist_tl],
+        grapheme,
+        [minimum, ..new_distlist],
+        minimum,
+      )
+    }
+  }
+}
+
+fn do_distance(a: String, b: String, distlist: List(Int), step: Int) {
+  case a {
+    "" -> result.unwrap(list.last(distlist), -1)
+    _ -> {
+      let assert Ok(src_hd) = string.first(a)
+      let src_tl = string.drop_left(a, up_to: 1)
+      let distlist = do_distlist(b, distlist, src_hd, [step], step)
+      do_distance(src_tl, b, distlist, step + 1)
+    }
+  }
+}
+
+fn levenshtein(a: String, b: String) -> Int {
+  case #(a, b) {
+    #(a, b) if a == b -> 0
+    #("", b) -> string.length(b)
+    #(a, "") -> string.length(a)
+    #(a, b) -> {
+      let distlist = list.range(0, string.length(b))
+      do_distance(a, b, distlist, 1)
+    }
+  }
+}
+
+fn unpercent(encoded: String) -> String {
+  let #([head], chunks) =
+    encoded
+    |> string.split(on: "%")
+    |> list.split(at: 1)
+
+  let assert Ok(res) =
+    chunks
+    |> list.map(fn(str) {
+      case string.length(str) < 2 {
+        True -> bit_string.from_string(str)
+        False -> {
+          let assert Ok(codepoint) =
+            str
+            |> string.slice(at_index: 0, length: 2)
+            |> int.base_parse(16)
+          <<codepoint:8, string.drop_left(str, 2):utf8>>
+        }
+      }
+    })
+    |> list.prepend(bit_string.from_string(head))
+    |> bit_string.concat
+    |> bit_string.to_string
+  res
+}
+
+fn guess_station(query: String, stations: Map(String, String)) -> String {
+  query
+  stations
+  |> map.keys
+  |> list.map(fn(a) { #(levenshtein(query, a), a) })
+  |> list.fold(
+    from: #(string.length(query), query),
+    with: fn(a, b) {
+      case a.0 < b.0 {
+        True -> a
+        False -> b
+      }
+    },
+  )
+  |> fn(a: #(Int, String)) { a.1 }
+}
+
 fn the_lookup(
   query: String,
   stations: Map(String, String),
@@ -21,7 +120,7 @@ fn the_lookup(
   |> lazy_unwrap(fn() {
     io.println(query)
     map.get(stations, query)
-    |> lazy_unwrap(fn() { "unknown" })
+    |> lazy_unwrap(fn() { guess_station(query, stations) })
   })
 }
 
@@ -32,13 +131,14 @@ fn lookup_station(
   baseurl: String,
 ) -> Response(BitBuilder) {
   let #(code, text) = case request {
-    Request(method: Get, path: "/help", ..) -> #(
+    Request(method: Get, path: "/help", ..)
+    | Request(method: Get, path: "/", ..) -> #(
       200,
-      "ds100 → Name: " <> baseurl <> "/FF\n" <> "Name → ds100: " <> baseurl <> "/Frankfurt Hbf",
+      "ds100 → Name: " <> baseurl <> "/NN\n" <> "Name → ds100: " <> baseurl <> "/Nürnberg Hbf",
     )
     Request(method: Get, path: "/" <> path, ..) -> #(
       200,
-      the_lookup(string.replace(path, each: "%20", with: " "), stations, ds100s),
+      the_lookup(unpercent(path), stations, ds100s),
     )
     _ -> #(404, "intended usage is e.g. curl " <> baseurl <> "/FF")
   }
@@ -49,6 +149,11 @@ fn lookup_station(
     "x-data-source",
     "https://data.deutschebahn.com/dataset/data-betriebsstellen.html",
   )
+  |> response.prepend_header(
+    "x-sources-at",
+    "https://stuebinm.eu/git/bahnhof.name",
+  )
+  |> response.prepend_header("content-type", "text/plain; charset=utf8")
   |> response.set_body(body)
 }
 
@@ -89,8 +194,8 @@ pub fn main() {
     |> list.map(fn(a) { #(a.1, a.0) })
     |> map.from_list
 
-  let assert Ok(_) = mist.run_service(
-    1234,
+  mist.run_service(
+    2345,
     fn(req) { lookup_station(req, stationmap, ds100map, baseurl) },
     max_body_limit: 100,
   )

test/bahnhofname_test.gleam

@@ -0,0 +1,12 @@
+import gleeunit
+import gleeunit/should
+
+pub fn main() {
+  gleeunit.main()
+}
+
+// gleeunit test functions end in `_test`
+pub fn hello_world_test() {
+  1
+  |> should.equal(1)
+}