From 00543b4e876cf069d5be22c1f06427a4c8d5bed9 Mon Sep 17 00:00:00 2001
From: Bruce Hill <bruce@bruce-hill.com>
Date: Wed, 4 Sep 2024 21:02:37 -0400
Subject: [PATCH] Add Text.map(pat, fn)

---
 builtins/text.c | 40 ++++++++++++++++++++++++++++++++++++++++
 builtins/text.h |  1 +
 docs/text.md    | 32 ++++++++++++++++++++++++++++++++
 environment.c   |  1 +
 test/text.tm    |  3 +++
 5 files changed, 77 insertions(+)

diff --git a/builtins/text.c b/builtins/text.c
index 59e12c6..97b50a4 100644
--- a/builtins/text.c
+++ b/builtins/text.c
@@ -1673,6 +1673,46 @@ public Text_t Text$replace(Text_t text, Pattern_t pattern, Text_t replacement, P
     return ret;
 }
 
+public Text_t Text$map(Text_t text, Pattern_t pattern, closure_t fn)
+{
+    Text_t ret = {.length=0};
+
+    int32_t first_grapheme = get_grapheme(pattern, 0);
+    bool find_first = (first_grapheme != '{'
+                       && !uc_is_property(first_grapheme, UC_PROPERTY_QUOTATION_MARK)
+                       && !uc_is_property(first_grapheme, UC_PROPERTY_PAIRED_PUNCTUATION));
+
+    iteration_state_t text_state = {0, 0};
+    int64_t nonmatching_pos = 0;
+
+    Text_t (*text_mapper)(Text_t, void*) = fn.fn;
+    for (int64_t pos = 0; pos < text.length; pos++) {
+        // Optimization: quickly skip ahead to first char in pattern:
+        if (find_first) {
+            while (pos < text.length && _next_grapheme(text, &text_state, pos) != first_grapheme)
+                ++pos;
+        }
+
+        int64_t match_len = match(text, pattern, pos, 0, NULL, 0);
+        if (match_len < 0) continue;
+
+        Text_t replacement = text_mapper(Text$slice(text, I(pos+1), I(pos+match_len)), fn.userdata);
+        if (pos > nonmatching_pos) {
+            Text_t before_slice = Text$slice(text, I(nonmatching_pos+1), I(pos));
+            ret = Text$concat(ret, before_slice, replacement);
+        } else {
+            ret = concat2(ret, replacement);
+        }
+        nonmatching_pos = pos + match_len;
+        pos += (match_len - 1);
+    }
+    if (nonmatching_pos < text.length) {
+        Text_t last_slice = Text$slice(text, I(nonmatching_pos+1), I(text.length));
+        ret = concat2(ret, last_slice);
+    }
+    return ret;
+}
+
 public Text_t Text$replace_all(Text_t text, table_t replacements, Text_t backref_pat, bool recursive)
 {
     if (replacements.entries.length == 0) return text;
diff --git a/builtins/text.h b/builtins/text.h
index a25a61d..c93c626 100644
--- a/builtins/text.h
+++ b/builtins/text.h
@@ -51,6 +51,7 @@ Text_t Text$from_codepoint_names(array_t codepoint_names);
 Text_t Text$from_bytes(array_t bytes);
 array_t Text$lines(Text_t text);
 Text_t Text$join(Text_t glue, array_t pieces);
+Text_t Text$map(Text_t text, Pattern_t pattern, closure_t fn);
 
 extern const TypeInfo $Text;
 
diff --git a/docs/text.md b/docs/text.md
index 0b02cdf..7039043 100644
--- a/docs/text.md
+++ b/docs/text.md
@@ -275,6 +275,7 @@ Text.has(pattern:Pattern)->Bool
 Text.find(pattern:Pattern, start=1, length=!&Int64?)->Int
 Text.find_all(pattern:Pattern)->[Text]
 Text.matches(pattern:Pattern)->Bool
+Text.map(pattern:Pattern, fn:func(t:Text)->Text)->Text
 Text.replace(pattern:Pattern, replacement:Text, placeholder:Pattern=$//)->[Text]
 Text.replace_all(replacements:{Pattern:Text}, placeholder:Pattern=$//)->[Text]
 Text.split(pattern:Pattern)->[Text]
@@ -837,6 +838,37 @@ matches(text: Text, pattern: Text) -> Bool
 
 ---
 
+## `map`
+
+**Description:**  
+For each occurrence of the given pattern, replace the text with the result of
+calling the given function on that text.
+
+**Usage:**  
+```tomo
+map(text: Text, pattern: Text, fn: func(text:Text)->Text) -> Text
+```
+
+**Parameters:**
+
+- `text`: The text to be searched.
+- `pattern`: The pattern to search for.
+- `fn`: The function to apply to each match.
+
+**Returns:**  
+The text with the matching parts replaced with the result of applying the given
+function to each.
+
+**Example:**  
+```tomo
+>> "hello world":map($/world/, Text.upper)
+= "hello WORLD"
+>> "Some nums: 1 2 3 4":map($/{int}/, func(i:Text): "$(Int.from_text(i) + 10)")
+= "Some nums: 11 12 13 14"
+```
+
+---
+
 ## `quoted`
 
 **Description:**  
diff --git a/environment.c b/environment.c
index 8f5de0d..2c1a3f7 100644
--- a/environment.c
+++ b/environment.c
@@ -256,6 +256,7 @@ env_t *new_compilation_unit(CORD *libname)
             {"join", "Text$join", "func(glue:Text, pieces:[Text])->Text"},
             {"lines", "Text$lines", "func(text:Text)->[Text]"},
             {"lower", "Text$lower", "func(text:Text)->Text"},
+            {"map", "Text$map", "func(text:Text, pattern:Pattern, fn:func(text:Text)->Text)->Text"},
             {"matches", "Text$matches", "func(text:Text, pattern:Pattern)->Bool"},
             {"quoted", "Text$quoted", "func(text:Text, color=no)->Text"},
             {"replace", "Text$replace", "func(text:Text, pattern:Pattern, replacement:Text, backref=$/\\/, recursive=yes)->Text"},
diff --git a/test/text.tm b/test/text.tm
index af03c3c..74a5341 100644
--- a/test/text.tm
+++ b/test/text.tm
@@ -250,3 +250,6 @@ func main():
 	= no
 	>> "Hello...":matches($/{id}/)
 	= no
+
+	>> "hello world":map($/world/, Text.upper)
+	= "hello WORLD"