From 417320b27c5633e3c52b9a9f1f24113a47783de8 Mon Sep 17 00:00:00 2001
From: "Elf M. Sternberg" <elf.sternberg@gmail.com>
Date: Tue, 13 Oct 2020 18:01:20 -0700
Subject: [PATCH] FEAT Parsing the content for references

This features all of the reference types that I commonly use,
including the ORG mode [[Title]], #CamelCase, #lisp-case, and #colon:case.
There are still edge cases around capitalization and the mixing of symbols
and numbers, and I'll have to hack on those until I'm satisfied.
---
 server/nm-store/src/reference_parser.rs | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/server/nm-store/src/reference_parser.rs b/server/nm-store/src/reference_parser.rs
index 0dbd19b..98c39f2 100644
--- a/server/nm-store/src/reference_parser.rs
+++ b/server/nm-store/src/reference_parser.rs
@@ -33,7 +33,7 @@ pub(crate) fn find_links(document: &str) -> Vec<String> {
 
     finder.iter_nodes(root, &|node| {
         lazy_static! {
-            static ref RE_REFERENCES: BytesRegex = BytesRegex::new(r"(\[\[([^\]]+)\]\]|(\#[\w\-]+))").unwrap();
+            static ref RE_REFERENCES: BytesRegex = BytesRegex::new(r"(\[\[([^\]]+)\]\]|(\#[:\w\-]+))").unwrap();
         }
 
         match &node.data.borrow().value {
@@ -55,7 +55,7 @@ fn recase(title: &str) -> String {
         static ref RE_PASS1: Regex = Regex::new(r"(?P<s>.)(?P<n>[A-Z][a-z]+)").unwrap();
         static ref RE_PASS2: Regex = Regex::new(r"(?P<s>[[:lower:]]|\d)(?P<n>[[:upper:]])").unwrap();
         static ref RE_PASS4: Regex = Regex::new(r"(?P<s>[a-z])(?P<n>\d)").unwrap();
-        static ref RE_PASS3: Regex = Regex::new(r"(_|-| )+").unwrap();
+        static ref RE_PASS3: Regex = Regex::new(r"(:|_|-| )+").unwrap();
     }
 
     // This should panic if misused, so... :-)
@@ -65,7 +65,7 @@ fn recase(title: &str) -> String {
     let pass = RE_PASS1.replace_all(&pass, "$s $n");
     let pass = RE_PASS4.replace_all(&pass, "$s $n");
     let pass = RE_PASS2.replace_all(&pass, "$s $n");
-    RE_PASS3.replace_all(&pass, " ").to_string()
+    RE_PASS3.replace_all(&pass, " ").trim().to_string()
 }
 
 fn build_page_titles(references: &Vec<String>) -> Vec<String> {
@@ -80,6 +80,7 @@ fn build_page_titles(references: &Vec<String>) -> Vec<String> {
                 _ => "".to_string(),
             }
         })
+        .filter(|s| s.len() > 0)
         .collect()
 }
 
@@ -96,7 +97,10 @@ mod tests {
 - #Test3
 - #TestFourAndSo
 - #Test-six-is-six
+- #recipe:entree
 - #
+- #-_-
+- #--Prefixed
 - [[]]
 
 But *[[Test Seven]]* isn't.  And *#Test_Eight____is_Messed-up*
@@ -110,10 +114,12 @@ Right? [[
             "Test 3",
             "Test Four And So",
             "Test six is six",
+            "recipe entree",
+            "Prefixed",
             "Test Seven",
             "Test Eight is Messed up",
             "Test Bite Me",
         ];
-        assert!(res.iter().eq(expected.iter()));
+        assert!(res.iter().eq(expected.iter()), "{:?}", res);
     }
 }