Fix #1, add depth arguments

8 files changed, 154 insertions, 61 deletions

diff --git a/Cargo.lock b/Cargo.lock
index 313fe53..55d0635 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -258,7 +258,7 @@ dependencies = [
 
 [[package]]
 name = "mktoc"
-version = "1.0.0"
+version = "1.1.0"
 dependencies = [
  "criterion 0.2.11 (registry+https://github.com/rust-lang/crates.io-index)",
  "regex 1.3.1 (registry+https://github.com/rust-lang/crates.io-index)",
diff --git a/Cargo.toml b/Cargo.toml
index 4e2714f..544900d 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -2,7 +2,7 @@
 name = "mktoc"
 description = "Generate Table of Contents from Markdown files"
 license = "MIT"
-version = "1.0.0"
+version = "1.1.0"
 authors = ["Kevin Gimbel <hallo@kevingimbel.com>"]
 edition = "2018"
 
diff --git a/README.md b/README.md
index 735d499..5c31d84 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,5 @@
 # `mktoc`
-> Blazingly fast Table of Content generator
+> Blazingly fast Markdown Table of Content generator
 
 ![](https://github.com/kevingimbel/mktoc/workflows/Clippy%20check/badge.svg)
 
@@ -16,11 +16,11 @@
 
 ## About
 
-`mktoc` parses markdown files and generates a Table Of Content linking all headlines up to heading level 6 deep.
+`mktoc` parses markdown files and generates a Table Of Content linking all headlines up to heading level 6 deep, or as specified by command line arguments. A start depth and maximum depth can be specified.
 
 ## Installation
 
-`mktoc` can be installed using Cargo, the Rust package manager, or by downloading a binary from GitHub.
+`mktoc` can be installed using Cargo, the Rust package manager.
 
 ### Cargo
 
@@ -30,20 +30,43 @@ $ cargo install mktoc
 
 ### Binary
 
-Download latest release from [https://github.com/kevingimbel/mktoc/releases](https://github.com/kevingimbel/mktoc/releases) and place it somewhere in your `PATH`, e.g. `/usr/local/bin`.
+Binaries are actually not available yet. If you know how releasing binaries with Rust can be implemented, please let me know!
 
 ## Usage
 
 Specify `--write` to overwrite the given file, otherwise the modified content is written to stdout.
 
 ```
-# mktoc [--write] <FILE>
+# mktoc [--write] [--max-depth|-M] [--min-depth|-m] <FILE>
 $ mktoc --write README.md
+$ mktoc --write -m 2 -M 4 README.md
 ```
 
 See `mktoc --help` for list of all arguments and flags.
 
+```
+mktoc 1.1.0
+
+USAGE:
+    mktoc [FLAGS] [OPTIONS] <file>
+
+FLAGS:
+    -h, --help       Prints help information
+    -V, --version    Prints version information
+    -w, --write      
+
+OPTIONS:
+    -M, --max-depth <max-depth>     [default: 6]
+    -m, --min-depth <min-depth>     [default: 1]
+
+ARGS:
+    <file> 
+```
  
 ## Performance
 
-`mktoc` is blazingly fast. Large files such as the README examples in `tests/files/` render in 0.009s (9ms) on average.
+`mktoc` is fast but can probably be even faster! Pull Requests and bug reports are appreciated!
+
+## License
+
+MIT, see LICENSE file.
\ No newline at end of file
diff --git a/mktoc.iml b/mktoc.iml
new file mode 100644
index 0000000..7fe828a
--- /dev/null
+++ b/mktoc.iml
@@ -0,0 +1,15 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<module type="RUST_MODULE" version="4">
+  <component name="NewModuleRootManager" inherit-compiler-output="true">
+    <exclude-output />
+    <content url="file://$MODULE_DIR$">
+      <sourceFolder url="file://$MODULE_DIR$/src" isTestSource="false" />
+      <sourceFolder url="file://$MODULE_DIR$/examples" isTestSource="false" />
+      <sourceFolder url="file://$MODULE_DIR$/tests" isTestSource="true" />
+      <sourceFolder url="file://$MODULE_DIR$/benches" isTestSource="true" />
+      <excludeFolder url="file://$MODULE_DIR$/target" />
+    </content>
+    <orderEntry type="inheritedJdk" />
+    <orderEntry type="sourceFolder" forTests="false" />
+  </component>
+</module>
\ No newline at end of file
diff --git a/src/bin.rs b/src/bin.rs
index 83638e6..29bba8b 100644
--- a/src/bin.rs
+++ b/src/bin.rs
@@ -8,6 +8,12 @@ struct Cli {
 
     #[structopt(long, short)]
     write: bool,
+
+    #[structopt(long, short="m", default_value="1")]
+    min_depth: i32,
+
+    #[structopt(long, short="M", default_value="6")]
+    max_depth: i32,
 }
 
 fn handle_write(new_toc: String) {
@@ -30,7 +36,7 @@ fn handle_write(new_toc: String) {
 
 fn main() {
     let opts = Cli::from_args();
-    let res = mktoc::make_toc(opts.file);
+    let res = mktoc::make_toc(opts.file, opts.min_depth, opts.max_depth);
 
     match res {
         Ok(new_toc) => {
diff --git a/src/lib.rs b/src/lib.rs
index c285417..ce4e8d3 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -16,6 +16,7 @@ struct Cli {
 const COMMENT_BEGIN: &str = "<!-- BEGIN mktoc -->";
 const COMMENT_END: &str = "<!-- END mktoc -->";
 
+/// reads a file into a mutable string
 fn read_file(file_path: String) -> Result<String, ::std::io::Error> {
     let mut file = File::open(file_path)?;
     let mut contents = String::new();
@@ -23,33 +24,62 @@ fn read_file(file_path: String) -> Result<String, ::std::io::Error> {
 
     Ok(contents)
 }
-
-fn generate_toc(original_content: String) -> String {
-    // @TODO: This RegEx creates weird outputs if the headline contains additional markdown, like images
-    let re = regex::Regex::new(r"((#{1,6}\s))((.*))").unwrap();
+/// parses a string and extracts all headlines to build a table of contents
+///
+/// Uses basic regex "((#{1,6}\s))((.*))" to parse headings. Right now this produces errors because it also matches
+/// comments in code blocks and if the headline contains images those will be matched, too.
+pub fn generate_toc(original_content: String, min_depth: i32, max_depth: i32) -> String {
+    let mut already_found_code_open = false;
+    let mut code_block_found = false;
     let mut new_toc = String::from(COMMENT_BEGIN);
-    // let caps = re.captures(content.as_str());
-    for caps in re.captures_iter(original_content.as_str()) {
-        let level: usize = caps.get(2).unwrap().as_str().chars().count() - 1;
-        let text = caps.get(3).unwrap().as_str();
-        // @TODO: Use real URL encoding
-        let link = text.replace(" ", "-").to_ascii_lowercase();
-        // let spaces = " ".repeat(level -1);
-        let spaces = match level {
-            3 => String::from("  "),
-            4 => String::from("    "),
-            5 => String::from("      "),
-            6 => String::from("        "),
-            _ => String::from(""),
-        };
-
-        new_toc = format!(
-            "{old}\n{spaces}- [{text}](#{link})",
-            old = new_toc,
-            spaces = spaces,
-            text = text,
-            link = link
-        );
+    let re = regex::Regex::new(r"((#{1,6}\s))((.*))").unwrap();
+    for line in original_content.lines() {
+
+        if line.starts_with("```") {
+            code_block_found = true;
+        }
+
+        if !code_block_found && !already_found_code_open {
+            if line.starts_with("#") {
+                let caps = re.captures(line).unwrap();
+                let level: i32 = (caps.get(2).unwrap().as_str().chars().count() - 1) as i32;
+                if level < min_depth {
+                    continue;
+                }
+
+                if level > max_depth {
+                    continue;
+                }
+
+
+                let text = caps.get(3).unwrap().as_str();
+                let link = text.replace(" ", "-").to_ascii_lowercase();
+                let spaces = match level {
+                    3 => String::from("  "),
+                    4 => String::from("    "),
+                    5 => String::from("      "),
+                    6 => String::from("        "),
+                    _ => String::from(""),
+                };
+                new_toc = format!(
+                    "{old}\n{spaces}- [{text}](#{link})",
+                    old = new_toc.as_str(),
+                    spaces = spaces,
+                    text = text,
+                    link = link
+                );
+            }
+        }
+
+        if code_block_found && already_found_code_open {
+            code_block_found = false;
+            already_found_code_open = false;
+        }
+
+        if line.starts_with("```") {
+            already_found_code_open = true;
+        }
+
     }
 
     new_toc = format!("{}\n{}", new_toc, COMMENT_END);
@@ -57,9 +87,10 @@ fn generate_toc(original_content: String) -> String {
     new_toc
 }
 
-pub fn make_toc(file_path_in: String) -> Result<String, ::std::io::Error> {
+/// takes a file path as `String` and returns a table of contents for the file
+pub fn make_toc(file_path_in: String, min_depth: i32, max_depth: i32) -> Result<String, ::std::io::Error> {
     let content = read_file(file_path_in)?;
-    let new_toc = generate_toc(content.to_owned());
+    let new_toc = generate_toc(content.to_owned(), min_depth, max_depth);
     let re_toc = regex::Regex::new(r"(?ms)^(<!-- BEGIN mktoc).*(END mktoc -->)").unwrap();
     let res: String = re_toc
         .replace_all(content.as_str(), new_toc.as_str())
diff --git a/tests/files/README_02.md b/tests/files/README_02.md
index 1fc07b2..be4a991 100644
--- a/tests/files/README_02.md
+++ b/tests/files/README_02.md
@@ -2,31 +2,25 @@ This README.md is taken from [https://github.com/terraform-aws-modules/terraform
 
 ---
 <!-- BEGIN mktoc -->
-- [AWS VPC Terraform module](#AWS-VPC-Terraform-module)
-- [Terraform versions](#Terraform-versions)
-- [Usage](#Usage)
-- [External NAT Gateway IPs](#External-NAT-Gateway-IPs)
-- [The rest of arguments are omitted for brevity](#The-rest-of-arguments-are-omitted-for-brevity)
-- [<= Skip creation of EIPs for the NAT Gateways](#<=-Skip-creation-of-EIPs-for-the-NAT-Gateways)
-- [<= IPs specified here as input to the module](#<=-IPs-specified-here-as-input-to-the-module)
-- [NAT Gateway Scenarios](#NAT-Gateway-Scenarios)
-  - [One NAT Gateway per subnet (default)](#One-NAT-Gateway-per-subnet-(default))
-  - [Single NAT Gateway](#Single-NAT-Gateway)
-  - [One NAT Gateway per availability zone](#One-NAT-Gateway-per-availability-zone)
+- [AWS VPC Terraform module](#aws-vpc-terraform-module)
+- [Terraform versions](#terraform-versions)
+- [Usage](#usage)
+- [External NAT Gateway IPs](#external-nat-gateway-ips)
+- [NAT Gateway Scenarios](#nat-gateway-scenarios)
+  - [One NAT Gateway per subnet (default)](#one-nat-gateway-per-subnet-(default))
+  - [Single NAT Gateway](#single-nat-gateway)
+  - [One NAT Gateway per availability zone](#one-nat-gateway-per-availability-zone)
 - ["private" versus "intra" subnets](#"private"-versus-"intra"-subnets)
-- [Conditional creation](#Conditional-creation)
-- [This VPC will not be created](#This-VPC-will-not-be-created)
-- [... omitted](#...-omitted)
-- [Public access to RDS instances](#Public-access-to-RDS-instances)
-- [Network Access Control Lists (ACL or NACL)](#Network-Access-Control-Lists-(ACL-or-NACL))
-- [Public access to Redshift cluster](#Public-access-to-Redshift-cluster)
-- [<= By default Redshift subnets will be associated with the private route table](#<=-By-default-Redshift-subnets-will-be-associated-with-the-private-route-table)
-- [Examples](#Examples)
-- [Inputs](#Inputs)
-- [Outputs](#Outputs)
-- [Tests](#Tests)
-- [Authors](#Authors)
-- [License](#License)
+- [Conditional creation](#conditional-creation)
+- [Public access to RDS instances](#public-access-to-rds-instances)
+- [Network Access Control Lists (ACL or NACL)](#network-access-control-lists-(acl-or-nacl))
+- [Public access to Redshift cluster](#public-access-to-redshift-cluster)
+- [Examples](#examples)
+- [Inputs](#inputs)
+- [Outputs](#outputs)
+- [Tests](#tests)
+- [Authors](#authors)
+- [License](#license)
 <!-- END mktoc -->
 
 <!-- START doctoc generated TOC please keep comment here to allow auto update -->
diff --git a/tests/files/README_04_code_block_issues.md b/tests/files/README_04_code_block_issues.md
new file mode 100644
index 0000000..c327a37
--- /dev/null
+++ b/tests/files/README_04_code_block_issues.md
@@ -0,0 +1,24 @@
+# Test file
+
+<!-- BEGIN mktoc -->
+- [Test for issue #1](#test-for-issue-#1)
+  - [Test Heading 3](#test-heading-3)
+    - [Test Heading 4](#test-heading-4)
+<!-- END mktoc -->
+
+## Test for issue #1
+
+```
+# This comment should not be added to the ToC
+function test(A,B) {
+  return A==B;
+}
+```
+
+### Test Heading 3
+
+#### Test Heading 4
+
+##### Test Heading 5
+
+###### Test Heading 6
\ No newline at end of file