From c19dd096c38cdaf7b9d3f642d1e31d44ac471531 Mon Sep 17 00:00:00 2001
From: Florian Schroedl <flo.schroedl@gmail.com>
Date: Thu, 20 Jan 2022 17:00:00 +0100
Subject: [PATCH] Document and move tryParseBuild

---
 src/org/org_text_delimiter.nim |  5 ++-
 src/org/org_types.nim          | 27 +++++++++++--
 src/parser/builder_api.nim     | 72 ++++++++++++++++++++++++++++++++++
 3 files changed, 99 insertions(+), 5 deletions(-)
 create mode 100644 src/parser/builder_api.nim

diff --git a/src/org/org_text_delimiter.nim b/src/org/org_text_delimiter.nim
index 43d07b2..63c7309 100644
--- a/src/org/org_text_delimiter.nim
+++ b/src/org/org_text_delimiter.nim
@@ -8,6 +8,7 @@ import ./org_types
 import ../utils/fp
 import ../parser/parser_internals
 import ../parser/parser_types
+import ../parser/builder_api
 
 let parseBetweenDelimiter* = proc(start: (Parser -> ParserResult), stop: (Parser -> ParserResult)): (Parser -> ParserResult) {.closure.} =
   ignore(start) +
@@ -74,7 +75,7 @@ proc makeOrgToken*(orgTokenFn: string -> OrgBuilderT): (seq[ParserToken], seq[Or
   return proc(parserTokens: seq[ParserToken], builderTokens: seq[OrgBuilderT]): seq[OrgBuilderT] =
     return builderTokens & parserTokens.foldl(a & b.tokenStringValue(), "").orgTokenFn()
 
-proc parseText[T](
+proc tryParseBuild[T](
   builder: Builder[T],
   builderFns: seq[tuple[
     parserFn: Parser -> ParserResult,
@@ -149,7 +150,7 @@ when isMainModule:
       parser: initParser("Regular *bold* /italic/ _underline_ =verbatim= ~code~ +strikethrough+"),
       tree: newSeq[OrgBuilderT](),
   )))
-  .flatMap((builder: OrgBuilder) => parseText(
+  .flatMap((builder: OrgBuilder) => tryParseBuild(
     builder = builder,
     builderFns = @[
       (boldParser, makeOrgToken(makeBoldToken)),
diff --git a/src/org/org_types.nim b/src/org/org_types.nim
index 19e8715..a0c5ac6 100644
--- a/src/org/org_types.nim
+++ b/src/org/org_types.nim
@@ -4,7 +4,14 @@ type
   orgElementKind* = enum
     orgRawText,
     orgText,
+
+    # Formating
     orgBoldText,
+    orgItalicText,
+    orgUnderlineText,
+    orgVerbatimText,
+    orgCodeText,
+    orgStrikeThroughText,
   OrgElement* = ref object
     children*: seq[OrgElement]
     content*: string
@@ -12,13 +19,27 @@ type
     case kind*: orgElementKind
     of orgRawText: discard
     of orgText: discard
+
+    # Formating
     of orgBoldText: discard
+    of orgItalicText: discard
+    of orgUnderlineText: discard
+    of orgVerbatimText: discard
+    of orgCodeText: discard
+    of orgStrikeThroughText: discard
 
 proc `$`*(x: orgElementKind): string =
   case x:
-    of orgRawText: "Text (Raw)"
-    of orgText: "Text"
-    of orgBoldText: "Text (Bold)"
+    of orgRawText:            "Text (Raw)"
+    of orgText:               "Text"
+
+    # Formating
+    of orgBoldText:           "Text (Bold)"
+    of orgItalicText:         "Text (Italic)"
+    of orgUnderlineText:      "Text (Underline)"
+    of orgVerbatimText:       "Text (Verbatim)"
+    of orgCodeText:           "Text (Code)"
+    of orgStrikeThroughText:  "Text (StrikeThrough)"
 
 proc `$`*(x: OrgElement): string =
   &"""OrgElement(
diff --git a/src/parser/builder_api.nim b/src/parser/builder_api.nim
new file mode 100644
index 0000000..69c0582
--- /dev/null
+++ b/src/parser/builder_api.nim
@@ -0,0 +1,72 @@
+import std/sugar
+import std/collections/sequtils
+import std/strformat
+import std/strutils
+import results
+import fusion/matching
+import ./parser_internals
+import ./parser_types
+
+proc tryParseBuild*[T](
+  builder: Builder[T],
+  builderFns: seq[tuple[
+    parserFn: Parser -> ParserResult,
+    concatFn: (seq[ParserToken], seq[T]) -> seq[T],
+  ]],
+  defaultBuilderFn: (seq[ParserToken]) -> seq[T],
+  stopAtParserFn = newline,
+): BuilderResult[T] =
+  ## Parse remaining text in `builder` by going checking in the `builderFns` list for a sucessful `parserFn`.
+  ## The `ok` `parserFn` result will be merged into the `Builder[T].tree` by using the `concatFn`.
+  ## Otherwise continue taking any character until the `stopAtParserFn` condition is found.
+  ## Any non-matching tokens will be converted using the `defaultBuilderFn`.
+  let (parser, tree) = builder
+
+  # Mutating accumulators
+  var parserAcc: ParserResult = ParserResult.ok(parser)
+  var builderAcc: Builder[T] = builder
+
+  while parserAcc.isOk() and parserAcc.flatMap(stopAtParserFn).isErr():
+    # Empty the parser tokens as we want to seperate them for the next parser in the sequence
+    let emptyParser = parserAcc.map(emptyTokens)
+
+    # Find the first matching parser and convert its tokens
+    var found = false
+    for builderFn in builderFns:
+      let (parserFn, concatFn) = builderFn
+
+      let parseResult = emptyParser.flatMap(parserFn)
+      if parseResult.isOk():
+        let okParser = parseResult.unsafeGet()
+
+        # Convert all previous unmatched tokens via the `defaultBuilderTokens`
+        let defaultBuilderTokens = parserAcc
+        .foldTokens(
+          onError = _ => newSeq[T](),
+          onSuccess = defaultBuilderFn,
+        )
+
+        found = true
+        parserAcc = parseResult.map(emptyTokens)
+        builderAcc = builder.initBuilder(
+          okParser,
+          concatFn(
+            okParser.tokens,
+            builderAcc[1] & defaultBuilderTokens,
+          ),
+        )
+        break
+
+    if not found:
+      parserAcc = parserAcc.flatMap(anyCh)
+
+  let defaultBuilderTokens = parserAcc
+  .foldTokens(
+    onError = _ => newSeq[T](),
+    onSuccess = defaultBuilderFn,
+  )
+
+  BuilderResult[T].ok(builder.initBuilder(
+    builderAcc[0],
+    builderAcc[1] & defaultBuilderTokens,
+  ))