From d02e44f82dea8095d53293100944b0361ad2110f Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Wed, 26 Nov 2025 16:44:02 -0500
Subject: [PATCH 01/13] Add --summary flag to cli output

---
 rust/cli/src/main.rs | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/rust/cli/src/main.rs b/rust/cli/src/main.rs
index 6f5c64bc..9d463c81 100644
--- a/rust/cli/src/main.rs
+++ b/rust/cli/src/main.rs
@@ -48,6 +48,10 @@ struct Flags {
     #[arg(long)]
     no_dereference: bool,
 
+    /// Prints a summary of file types at the end of the output.
+    #[arg(long)]
+    summary: bool,
+
     #[clap(flatten)]
     colors: Colors,
 
@@ -211,12 +215,26 @@ async fn main() -> Result<()> {
     if flags.format.json {
         print!("[");
     }
+    // Initialize counter for file types
+    let mut type_counts: HashMap<String, usize> = HashMap::new();
+    // Reorder responses to match input order
     let mut reorder = Reorder::default();
     let mut errors = false;
     while let Some(response) = result_receiver.recv().await {
         reorder.push(response?);
         while let Some(response) = reorder.pop() {
             errors |= response.result.is_err();
+            // Count file type for summary
+            if flags.summary {
+                // If result is ok, extract the description (e.g., "Python source")
+                if let Ok(file_type) = &response.result {
+                    // Can change .description to .label for shorter names like "python"
+                    let label = file_type.info().description.to_string();
+                    // If the label exists, add 1. If not, insert 0 then add 1.
+                    *type_counts.entry(label).or_insert(0) += 1;
+                }
+            }
+            // Print output
             if flags.format.json {
                 if reorder.next != 1 {
                     print!(",");
@@ -236,6 +254,16 @@ async fn main() -> Result<()> {
         }
         println!("]");
     }
+    // Print summary if requested
+    if flags.summary && !flags.format.json && !flags.format.jsonl {
+        println!("--- Summary ---");
+        // Sort the results by count (descending)
+        let mut sorted_counts: Vec<_> = type_counts.into_iter().collect();
+        sorted_counts.sort_by(|a, b| b.1.cmp(&a.1));
+        for (label, count) in sorted_counts {
+            println!("{}: {}", label, count);
+        }
+    }
     if errors {
         std::process::exit(1);
     }

From d735809723f90007f8c67fcda21fd4caa9a7dcf0 Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Wed, 26 Nov 2025 17:05:25 -0500
Subject: [PATCH 02/13] Update summary to only display if no errors

---
 rust/cli/src/main.rs | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/rust/cli/src/main.rs b/rust/cli/src/main.rs
index 9d463c81..335bdeef 100644
--- a/rust/cli/src/main.rs
+++ b/rust/cli/src/main.rs
@@ -254,7 +254,10 @@ async fn main() -> Result<()> {
         }
         println!("]");
     }
-    // Print summary if requested
+    if errors {
+        std::process::exit(1);
+    }
+    // Print summary if requested, only if there were no errors
     if flags.summary && !flags.format.json && !flags.format.jsonl {
         println!("--- Summary ---");
         // Sort the results by count (descending)
@@ -264,9 +267,6 @@ async fn main() -> Result<()> {
             println!("{}: {}", label, count);
         }
     }
-    if errors {
-        std::process::exit(1);
-    }
     Ok(())
 }
 

From 871e8fac1953fc365d1a148caabc0565593a3a40 Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Sat, 29 Nov 2025 06:09:31 -0500
Subject: [PATCH 03/13] Add color and alphabetical sorting to summary

---
 rust/cli/src/main.rs | 55 +++++++++++++++++++++++++++++++-------------
 1 file changed, 39 insertions(+), 16 deletions(-)

diff --git a/rust/cli/src/main.rs b/rust/cli/src/main.rs
index 335bdeef..09481b72 100644
--- a/rust/cli/src/main.rs
+++ b/rust/cli/src/main.rs
@@ -22,6 +22,7 @@ use std::sync::Arc;
 use anyhow::{bail, ensure, Result};
 use clap::{Args, Parser};
 use colored::ColoredString;
+use colored::Colorize;
 use magika_lib::{
     self as magika, ContentType, Features, FeaturesOrRuled, FileType, InferredType,
     OverwriteReason, Session, TypeInfo,
@@ -215,26 +216,26 @@ async fn main() -> Result<()> {
     if flags.format.json {
         print!("[");
     }
-    // Initialize counter for file types
-    let mut type_counts: HashMap<String, usize> = HashMap::new();
-    // Reorder responses to match input order
+    // Initializes counter for file types.
+    let mut type_counts: HashMap<(String, String), usize> = HashMap::new();
     let mut reorder = Reorder::default();
     let mut errors = false;
+    // Prints results, reordering as needed to match input order.
     while let Some(response) = result_receiver.recv().await {
         reorder.push(response?);
         while let Some(response) = reorder.pop() {
             errors |= response.result.is_err();
-            // Count file type for summary
+            // Counts file type for the final summary.
             if flags.summary {
-                // If result is ok, extract the description (e.g., "Python source")
+                // If result is ok, extracts the description (e.g., "Python source").
                 if let Ok(file_type) = &response.result {
-                    // Can change .description to .label for shorter names like "python"
-                    let label = file_type.info().description.to_string();
-                    // If the label exists, add 1. If not, insert 0 then add 1.
-                    *type_counts.entry(label).or_insert(0) += 1;
+                    let type_label = file_type.info().description.to_string();
+                    let group = file_type.info().group.to_string();
+                    // Increments the count in the HashMap, inserting 0 if the key does not exist.
+                    *type_counts.entry((type_label, group)).or_insert(0) += 1;
                 }
             }
-            // Print output
+            // Prints output.
             if flags.format.json {
                 if reorder.next != 1 {
                     print!(",");
@@ -257,14 +258,23 @@ async fn main() -> Result<()> {
     if errors {
         std::process::exit(1);
     }
-    // Print summary if requested, only if there were no errors
+    // Prints summary if requested (only if there were no errors).
     if flags.summary && !flags.format.json && !flags.format.jsonl {
         println!("--- Summary ---");
-        // Sort the results by count (descending)
+        // Sorts by count (descending).
         let mut sorted_counts: Vec<_> = type_counts.into_iter().collect();
-        sorted_counts.sort_by(|a, b| b.1.cmp(&a.1));
-        for (label, count) in sorted_counts {
-            println!("{}: {}", label, count);
+        sorted_counts.sort_by(|a, b| {
+            let count_cmp = b.1.cmp(&a.1);
+            if count_cmp == std::cmp::Ordering::Equal {
+                // Sorts alphabetically (case-insensitive).
+                // a.0 .0 accesses the type_label.
+                a.0 .0.to_lowercase().cmp(&b.0 .0.to_lowercase())
+            } else {
+                count_cmp
+            }
+        });
+        for ((type_label, group), count) in sorted_counts {
+            println!("{}: {}", color_type_label(&type_label, &group), count);
         }
     }
     Ok(())
@@ -582,7 +592,6 @@ impl Response {
     }
 
     fn color(&self, result: ColoredString) -> ColoredString {
-        use colored::Colorize as _;
         // We only use true colors (except for errors). If the terminal doesn't support true colors,
         // the colored crate will automatically choose the closest one.
         match &self.result {
@@ -615,3 +624,17 @@ fn join<T: AsRef<str>>(xs: impl IntoIterator<Item = T>) -> String {
     result.push(']');
     result
 }
+
+fn color_type_label(type_label: &str, group: &str) -> colored::ColoredString {
+    match group {
+        "application" => type_label.truecolor(0xf4, 0x3f, 0x5e),
+        "archive" => type_label.truecolor(0xf5, 0x9e, 0x0b),
+        "audio" => type_label.truecolor(0x84, 0xcc, 0x16),
+        "code" => type_label.truecolor(0x8b, 0x5c, 0xf6),
+        "document" => type_label.truecolor(0x3b, 0x82, 0xf6),
+        "executable" => type_label.truecolor(0xec, 0x48, 0x99),
+        "image" => type_label.truecolor(0x06, 0xb6, 0xd4),
+        "video" => type_label.truecolor(0x10, 0xb9, 0x81),
+        _ => type_label.bold().truecolor(0xcc, 0xcc, 0xcc),
+    }
+}

From 8ff868ddbb1d54951338b367d91fca7afd8c6d5e Mon Sep 17 00:00:00 2001
From: Ricardo <rjbarrioslll@gmail.com>
Date: Sat, 29 Nov 2025 22:12:57 -0500
Subject: [PATCH 04/13] Implemented scan_directory and added necessary testing
 files

---
 python/src/magika/magika.py               | 27 +++++++++++++++++
 python/tests/test_magika_python_module.py |  6 ++++
 python/tests/utils.py                     |  6 ++++
 test_magika.py                            | 24 +++++++++++++++
 tests_data/scan_directory/code2.py        | 32 ++++++++++++++++++++
 tests_data/scan_directory/test.json       | 31 +++++++++++++++++++
 tests_data/scan_directory/test.rs         | 36 +++++++++++++++++++++++
 tests_data/scan_directory/test.ts         | 36 +++++++++++++++++++++++
 8 files changed, 198 insertions(+)
 create mode 100644 test_magika.py
 create mode 100644 tests_data/scan_directory/code2.py
 create mode 100644 tests_data/scan_directory/test.json
 create mode 100644 tests_data/scan_directory/test.rs
 create mode 100644 tests_data/scan_directory/test.ts

diff --git a/python/src/magika/magika.py b/python/src/magika/magika.py
index fa2152ea..ece293c1 100644
--- a/python/src/magika/magika.py
+++ b/python/src/magika/magika.py
@@ -167,6 +167,33 @@ def identify_paths(
 
         return self._get_results_from_paths(paths_)
 
+    def scan_directory(self, directory: Union[str, os.PathLike]) -> List[MagikaResult]:
+        """Identify the content type of a all files in a directory given its path."""
+        path_obj = Path(directory)
+        
+        # Guard clause: check if directory exists
+        if not path_obj.exists() or not path_obj.is_dir():
+            raise FileNotFoundError(f"The directory '{directory}' does not exist or is not a directory.")
+
+        collected_paths: List[Union[str, os.PathLike]] = []
+
+        # rglob('*') recursively finds all files and directories
+        for item in path_obj.rglob('*'):
+            # We only want files, not sub-directories themselves
+            if item.is_file():
+                collected_paths.append(item)
+
+        paths_ = []
+        for path in collected_paths:
+            if isinstance(path, str) or isinstance(path, os.PathLike):
+                paths_.append(Path(path))
+            else:
+                raise TypeError(
+                    f"Input '{path}' is invalid: input path should be of type `Union[str, os.PathLike]`"
+                )
+
+        return self._get_results_from_paths(paths_)
+
     def identify_bytes(self, content: bytes) -> MagikaResult:
         """Identify the content type of raw bytes."""
         if not isinstance(content, bytes):
diff --git a/python/tests/test_magika_python_module.py b/python/tests/test_magika_python_module.py
index e25bd40b..c7595e64 100644
--- a/python/tests/test_magika_python_module.py
+++ b/python/tests/test_magika_python_module.py
@@ -72,6 +72,12 @@ def test_magika_module_with_explicit_model_dir() -> None:
     with open(test_path, "rb") as f:
         _ = m.identify_stream(f)
 
+def test_magika_module_with_basic_tests_by_directory() -> None:
+    tests_paths = utils.get_directory_test_dir()
+
+    m = Magika()
+    results = m.scan_directory(tests_paths)
+
 
 def test_magika_module_with_basic_tests_by_paths() -> None:
     tests_paths = utils.get_basic_test_files_paths()
diff --git a/python/tests/utils.py b/python/tests/utils.py
index 375dcd8c..8a38eac1 100644
--- a/python/tests/utils.py
+++ b/python/tests/utils.py
@@ -42,6 +42,12 @@ def get_basic_tests_files_dir() -> Path:
     return tests_files_dir
 
 
+def get_directory_test_dir() -> Path:
+    tests_files_dir = get_tests_data_dir() / "scan_directory"
+    assert tests_files_dir.is_dir()
+    return tests_files_dir
+
+
 def get_mitra_tests_files_dir() -> Path:
     tests_files_dir = get_tests_data_dir() / "mitra"
     assert tests_files_dir.is_dir()
diff --git a/test_magika.py b/test_magika.py
new file mode 100644
index 00000000..a5272fd3
--- /dev/null
+++ b/test_magika.py
@@ -0,0 +1,24 @@
+#from magika import Magika
+import sys
+import os
+
+# Get the absolute path to the folder containing magika.py
+# We go: current dir -> python -> src -> magika
+module_path = os.path.abspath(os.path.join(os.path.dirname(__file__), 'python', 'src'))
+
+# --- THE FIX ---
+# Use insert(0, path) instead of append(path)
+# This puts your folder at the TOP of the search list
+if module_path not in sys.path:
+    sys.path.insert(0, module_path)
+
+import magika
+#m = Magika()
+m = magika.Magika()
+#res = m.identify_path('./tests_data/basic/python')
+res = m.scan_directory('./tests_data/scan_directory')
+print("------------------ PREDICTIONS ------------------")
+for result in res:
+    print(result)
+    print()
+#print(res)
\ No newline at end of file
diff --git a/tests_data/scan_directory/code2.py b/tests_data/scan_directory/code2.py
new file mode 100644
index 00000000..a2931d6d
--- /dev/null
+++ b/tests_data/scan_directory/code2.py
@@ -0,0 +1,32 @@
+import random
+import time
+
+class NumberGuesser:
+    def __init__(self, limit=100):
+        self.limit = limit
+        self.target = random.randint(1, limit)
+        self.attempts = 0
+
+    def guess(self, user_input):
+        self.attempts += 1
+        try:
+            val = int(user_input)
+        except ValueError:
+            return "Please enter a valid integer."
+
+        if val < self.target:
+            return "Too low!"
+        elif val > self.target:
+            return "Too high!"
+        else:
+            return f"Correct! It took you {self.attempts} tries."
+
+def main():
+    game = NumberGuesser()
+    print("I'm thinking of a number between 1 and 100.")
+    # Simulation of a game loop
+    for i in range(5):
+        print(f"Simulation guess {i}: {game.guess(i * 20)}")
+
+if __name__ == "__main__":
+    main()
\ No newline at end of file
diff --git a/tests_data/scan_directory/test.json b/tests_data/scan_directory/test.json
new file mode 100644
index 00000000..df29c36a
--- /dev/null
+++ b/tests_data/scan_directory/test.json
@@ -0,0 +1,31 @@
+{
+  "project": {
+    "id": 1024,
+    "name": "MagikaTest",
+    "isActive": true,
+    "tags": ["python", "cpp", "automation"],
+    "metadata": {
+      "created_at": "2023-10-27T10:00:00Z",
+      "author": "dev_user",
+      "version": 1.2
+    }
+  },
+  "employees": [
+    {
+      "id": 1,
+      "name": "Alice Smith",
+      "role": "Lead Developer",
+      "skills": ["Rust", "C++"]
+    },
+    {
+      "id": 2,
+      "name": "Bob Jones",
+      "role": "QA Engineer",
+      "skills": ["Python", "Selenium"]
+    }
+  ],
+  "settings": {
+    "retry_attempts": 5,
+    "timeout_ms": 3000
+  }
+}
\ No newline at end of file
diff --git a/tests_data/scan_directory/test.rs b/tests_data/scan_directory/test.rs
new file mode 100644
index 00000000..3ea8704d
--- /dev/null
+++ b/tests_data/scan_directory/test.rs
@@ -0,0 +1,36 @@
+use std::collections::HashMap;
+
+struct User {
+    username: String,
+    email: String,
+    sign_in_count: u64,
+    active: bool,
+}
+
+impl User {
+    fn new(username: String, email: String) -> User {
+        User {
+            username,
+            email,
+            sign_in_count: 1,
+            active: true,
+        }
+    }
+
+    fn deactivate(&mut self) {
+        self.active = false;
+        println!("User {} has been deactivated.", self.username);
+    }
+}
+
+fn main() {
+    let mut users = HashMap::new();
+    let user1 = User::new(String::from("rust_fan"), String::from("rust@example.com"));
+    
+    users.insert("u1", user1);
+
+    match users.get_mut("u1") {
+        Some(u) => u.deactivate(),
+        None => println!("User not found"),
+    }
+}
\ No newline at end of file
diff --git a/tests_data/scan_directory/test.ts b/tests_data/scan_directory/test.ts
new file mode 100644
index 00000000..5773bc3a
--- /dev/null
+++ b/tests_data/scan_directory/test.ts
@@ -0,0 +1,36 @@
+interface Task {
+    id: number;
+    title: string;
+    completed: boolean;
+    completedAt?: Date; // Optional property
+}
+
+class TaskManager {
+    private tasks: Task[] = [];
+
+    addTask(title: string): void {
+        const newTask: Task = {
+            id: this.tasks.length + 1,
+            title: title,
+            completed: false
+        };
+        this.tasks.push(newTask);
+    }
+
+    completeTask(id: number): void {
+        const task = this.tasks.find(t => t.id === id);
+        if (task) {
+            task.completed = true;
+            task.completedAt = new Date();
+            console.log(`Task '${task.title}' marked as done.`);
+        }
+    }
+
+    listTasks(): Task[] {
+        return this.tasks;
+    }
+}
+
+const manager = new TaskManager();
+manager.addTask("Learn TypeScript");
+manager.completeTask(1);
\ No newline at end of file

From 57843b130a5671bdf6469a653d94e0626757f99e Mon Sep 17 00:00:00 2001
From: Ricardo <rjbarrioslll@gmail.com>
Date: Sat, 29 Nov 2025 22:23:40 -0500
Subject: [PATCH 05/13] cleanup

---
 test_magika.py | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/test_magika.py b/test_magika.py
index a5272fd3..301779a0 100644
--- a/test_magika.py
+++ b/test_magika.py
@@ -2,13 +2,8 @@
 import sys
 import os
 
-# Get the absolute path to the folder containing magika.py
-# We go: current dir -> python -> src -> magika
 module_path = os.path.abspath(os.path.join(os.path.dirname(__file__), 'python', 'src'))
 
-# --- THE FIX ---
-# Use insert(0, path) instead of append(path)
-# This puts your folder at the TOP of the search list
 if module_path not in sys.path:
     sys.path.insert(0, module_path)
 

From 23602718124dc4fcd56403c7278f34e32efc6822 Mon Sep 17 00:00:00 2001
From: Christian Lopez <christian@Christians-MacBook-Air-2.local>
Date: Sun, 30 Nov 2025 17:23:01 -0400
Subject: [PATCH 06/13] added recursive scanning to scan_directory

---
 python/src/magika/magika.py | 12 +++++++-----
 test_magika.py              | 19 -------------------
 2 files changed, 7 insertions(+), 24 deletions(-)
 delete mode 100644 test_magika.py

diff --git a/python/src/magika/magika.py b/python/src/magika/magika.py
index ece293c1..fb1ba574 100644
--- a/python/src/magika/magika.py
+++ b/python/src/magika/magika.py
@@ -166,9 +166,9 @@ def identify_paths(
                 )
 
         return self._get_results_from_paths(paths_)
-
-    def scan_directory(self, directory: Union[str, os.PathLike]) -> List[MagikaResult]:
-        """Identify the content type of a all files in a directory given its path."""
+    
+    def scan_directory(self, directory: Union[str, os.PathLike], recursive_scan=False) -> List[MagikaResult]:
+        """Identify the content type of all files in a directory given its path."""
         path_obj = Path(directory)
         
         # Guard clause: check if directory exists
@@ -177,8 +177,10 @@ def scan_directory(self, directory: Union[str, os.PathLike]) -> List[MagikaResul
 
         collected_paths: List[Union[str, os.PathLike]] = []
 
-        # rglob('*') recursively finds all files and directories
-        for item in path_obj.rglob('*'):
+        # Use rglob('*') for recursive scan, glob('*') for single directory
+        glob_pattern = path_obj.rglob('*') if recursive_scan else path_obj.glob('*')
+        
+        for item in glob_pattern:
             # We only want files, not sub-directories themselves
             if item.is_file():
                 collected_paths.append(item)
diff --git a/test_magika.py b/test_magika.py
deleted file mode 100644
index 301779a0..00000000
--- a/test_magika.py
+++ /dev/null
@@ -1,19 +0,0 @@
-#from magika import Magika
-import sys
-import os
-
-module_path = os.path.abspath(os.path.join(os.path.dirname(__file__), 'python', 'src'))
-
-if module_path not in sys.path:
-    sys.path.insert(0, module_path)
-
-import magika
-#m = Magika()
-m = magika.Magika()
-#res = m.identify_path('./tests_data/basic/python')
-res = m.scan_directory('./tests_data/scan_directory')
-print("------------------ PREDICTIONS ------------------")
-for result in res:
-    print(result)
-    print()
-#print(res)
\ No newline at end of file

From a5c0ab045788a49c7cbc89f330d08b8a77d363dd Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Mon, 1 Dec 2025 04:01:39 -0500
Subject: [PATCH 07/13] Fix format for nightly rust

---
 rust/cli/src/main.rs | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/rust/cli/src/main.rs b/rust/cli/src/main.rs
index 09481b72..fdb1a9aa 100644
--- a/rust/cli/src/main.rs
+++ b/rust/cli/src/main.rs
@@ -21,8 +21,7 @@ use std::sync::Arc;
 
 use anyhow::{bail, ensure, Result};
 use clap::{Args, Parser};
-use colored::ColoredString;
-use colored::Colorize;
+use colored::{ColoredString, Colorize};
 use magika_lib::{
     self as magika, ContentType, Features, FeaturesOrRuled, FileType, InferredType,
     OverwriteReason, Session, TypeInfo,

From 11eef79234fa6808a6ca5bea48d7c108818060fd Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Mon, 1 Dec 2025 04:15:12 -0500
Subject: [PATCH 08/13] Sync generated rust file to fix CI error

---
 rust/cli/README.md | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/rust/cli/README.md b/rust/cli/README.md
index ebaf28e7..f0dcb3d1 100644
--- a/rust/cli/README.md
+++ b/rust/cli/README.md
@@ -142,6 +142,9 @@ Options:
       --no-dereference
           Identifies symbolic links as is instead of identifying their content by following them
 
+      --summary
+          Prints a summary of file types at the end of the output
+
       --colors
           Prints with colors regardless of terminal support
 

From 6b451b0872d11d5f4710076a549e1c20ef852ff4 Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Mon, 1 Dec 2025 16:53:31 -0500
Subject: [PATCH 09/13] Fix syntax of scan_directory test

---
 python/tests/test_magika_python_module.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/python/tests/test_magika_python_module.py b/python/tests/test_magika_python_module.py
index c7595e64..a1d6fd57 100644
--- a/python/tests/test_magika_python_module.py
+++ b/python/tests/test_magika_python_module.py
@@ -76,7 +76,7 @@ def test_magika_module_with_basic_tests_by_directory() -> None:
     tests_paths = utils.get_directory_test_dir()
 
     m = Magika()
-    results = m.scan_directory(tests_paths)
+    _ = m.scan_directory(tests_paths)
 
 
 def test_magika_module_with_basic_tests_by_paths() -> None:

From b7978e205f862626c5212762af63f0e4d69daabc Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Mon, 1 Dec 2025 17:15:00 -0500
Subject: [PATCH 10/13] Fix format with ruff

---
 python/src/magika/magika.py               | 16 ++++++++++------
 python/tests/test_magika_python_module.py |  1 +
 2 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/python/src/magika/magika.py b/python/src/magika/magika.py
index fb1ba574..081ab4ce 100644
--- a/python/src/magika/magika.py
+++ b/python/src/magika/magika.py
@@ -166,20 +166,24 @@ def identify_paths(
                 )
 
         return self._get_results_from_paths(paths_)
-    
-    def scan_directory(self, directory: Union[str, os.PathLike], recursive_scan=False) -> List[MagikaResult]:
+
+    def scan_directory(
+        self, directory: Union[str, os.PathLike], recursive_scan=False
+    ) -> List[MagikaResult]:
         """Identify the content type of all files in a directory given its path."""
         path_obj = Path(directory)
-        
+
         # Guard clause: check if directory exists
         if not path_obj.exists() or not path_obj.is_dir():
-            raise FileNotFoundError(f"The directory '{directory}' does not exist or is not a directory.")
+            raise FileNotFoundError(
+                f"The directory '{directory}' does not exist or is not a directory."
+            )
 
         collected_paths: List[Union[str, os.PathLike]] = []
 
         # Use rglob('*') for recursive scan, glob('*') for single directory
-        glob_pattern = path_obj.rglob('*') if recursive_scan else path_obj.glob('*')
-        
+        glob_pattern = path_obj.rglob("*") if recursive_scan else path_obj.glob("*")
+
         for item in glob_pattern:
             # We only want files, not sub-directories themselves
             if item.is_file():
diff --git a/python/tests/test_magika_python_module.py b/python/tests/test_magika_python_module.py
index a1d6fd57..d31b7dfa 100644
--- a/python/tests/test_magika_python_module.py
+++ b/python/tests/test_magika_python_module.py
@@ -72,6 +72,7 @@ def test_magika_module_with_explicit_model_dir() -> None:
     with open(test_path, "rb") as f:
         _ = m.identify_stream(f)
 
+
 def test_magika_module_with_basic_tests_by_directory() -> None:
     tests_paths = utils.get_directory_test_dir()
 

From fc502c7be5293a7bc088c3ed2d87e1c511edc881 Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Mon, 1 Dec 2025 17:19:59 -0500
Subject: [PATCH 11/13] Fix missing type annotation

---
 python/src/magika/magika.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/python/src/magika/magika.py b/python/src/magika/magika.py
index 081ab4ce..120efad4 100644
--- a/python/src/magika/magika.py
+++ b/python/src/magika/magika.py
@@ -168,7 +168,7 @@ def identify_paths(
         return self._get_results_from_paths(paths_)
 
     def scan_directory(
-        self, directory: Union[str, os.PathLike], recursive_scan=False
+        self, directory: Union[str, os.PathLike], recursive_scan: bool = False
     ) -> List[MagikaResult]:
         """Identify the content type of all files in a directory given its path."""
         path_obj = Path(directory)

From f8322bfea3445f44e4b30cc3eeddb5575fa8830e Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Tue, 2 Dec 2025 00:40:26 -0500
Subject: [PATCH 12/13] Fix incomplete unit tests for scan_directory

---
 python/src/magika/magika.py                   |  2 +-
 python/tests/test_magika_python_module.py     | 14 ++++++--
 python/tests/utils.py                         |  4 +--
 tests_data/directory/txt/complex-sentence.txt |  1 +
 tests_data/scan_directory/code2.py            | 32 -----------------
 tests_data/scan_directory/test.json           | 31 ----------------
 tests_data/scan_directory/test.rs             | 36 -------------------
 tests_data/scan_directory/test.ts             | 36 -------------------
 8 files changed, 16 insertions(+), 140 deletions(-)
 create mode 100644 tests_data/directory/txt/complex-sentence.txt
 delete mode 100644 tests_data/scan_directory/code2.py
 delete mode 100644 tests_data/scan_directory/test.json
 delete mode 100644 tests_data/scan_directory/test.rs
 delete mode 100644 tests_data/scan_directory/test.ts

diff --git a/python/src/magika/magika.py b/python/src/magika/magika.py
index 120efad4..0ee892c9 100644
--- a/python/src/magika/magika.py
+++ b/python/src/magika/magika.py
@@ -182,7 +182,7 @@ def scan_directory(
         collected_paths: List[Union[str, os.PathLike]] = []
 
         # Use rglob('*') for recursive scan, glob('*') for single directory
-        glob_pattern = path_obj.rglob("*") if recursive_scan else path_obj.glob("*")
+        glob_pattern = sorted(path_obj.rglob("*")) if recursive_scan else sorted(path_obj.glob("*"))
 
         for item in glob_pattern:
             # We only want files, not sub-directories themselves
diff --git a/python/tests/test_magika_python_module.py b/python/tests/test_magika_python_module.py
index d31b7dfa..2ac1aa61 100644
--- a/python/tests/test_magika_python_module.py
+++ b/python/tests/test_magika_python_module.py
@@ -74,10 +74,20 @@ def test_magika_module_with_explicit_model_dir() -> None:
 
 
 def test_magika_module_with_basic_tests_by_directory() -> None:
-    tests_paths = utils.get_directory_test_dir()
+    tests_paths = utils.get_directory_tests_files_dir()
 
     m = Magika()
-    _ = m.scan_directory(tests_paths)
+
+    # Only scan direct children of tests_data/directory.
+    # Expected output is "directory" content type.
+    results = m.scan_directory(tests_paths)
+    direct_children = sorted([p for p in tests_paths.glob("*")])
+    check_results_vs_expected_results(direct_children, results)
+
+    # Scan all files recursively. Expected output is content type of each file.
+    results = m.scan_directory(tests_paths, recursive_scan=True)
+    all_files = sorted([p for p in tests_paths.rglob("*") if p.is_file()])
+    check_results_vs_expected_results(all_files, results)
 
 
 def test_magika_module_with_basic_tests_by_paths() -> None:
diff --git a/python/tests/utils.py b/python/tests/utils.py
index 8a38eac1..4796b20e 100644
--- a/python/tests/utils.py
+++ b/python/tests/utils.py
@@ -42,8 +42,8 @@ def get_basic_tests_files_dir() -> Path:
     return tests_files_dir
 
 
-def get_directory_test_dir() -> Path:
-    tests_files_dir = get_tests_data_dir() / "scan_directory"
+def get_directory_tests_files_dir() -> Path:
+    tests_files_dir = get_tests_data_dir() / "directory"
     assert tests_files_dir.is_dir()
     return tests_files_dir
 
diff --git a/tests_data/directory/txt/complex-sentence.txt b/tests_data/directory/txt/complex-sentence.txt
new file mode 100644
index 00000000..9b6a0c83
--- /dev/null
+++ b/tests_data/directory/txt/complex-sentence.txt
@@ -0,0 +1 @@
+This is yet another simple test, it includes one simple sentence, but it is not as trivial as other simpler tests.
\ No newline at end of file
diff --git a/tests_data/scan_directory/code2.py b/tests_data/scan_directory/code2.py
deleted file mode 100644
index a2931d6d..00000000
--- a/tests_data/scan_directory/code2.py
+++ /dev/null
@@ -1,32 +0,0 @@
-import random
-import time
-
-class NumberGuesser:
-    def __init__(self, limit=100):
-        self.limit = limit
-        self.target = random.randint(1, limit)
-        self.attempts = 0
-
-    def guess(self, user_input):
-        self.attempts += 1
-        try:
-            val = int(user_input)
-        except ValueError:
-            return "Please enter a valid integer."
-
-        if val < self.target:
-            return "Too low!"
-        elif val > self.target:
-            return "Too high!"
-        else:
-            return f"Correct! It took you {self.attempts} tries."
-
-def main():
-    game = NumberGuesser()
-    print("I'm thinking of a number between 1 and 100.")
-    # Simulation of a game loop
-    for i in range(5):
-        print(f"Simulation guess {i}: {game.guess(i * 20)}")
-
-if __name__ == "__main__":
-    main()
\ No newline at end of file
diff --git a/tests_data/scan_directory/test.json b/tests_data/scan_directory/test.json
deleted file mode 100644
index df29c36a..00000000
--- a/tests_data/scan_directory/test.json
+++ /dev/null
@@ -1,31 +0,0 @@
-{
-  "project": {
-    "id": 1024,
-    "name": "MagikaTest",
-    "isActive": true,
-    "tags": ["python", "cpp", "automation"],
-    "metadata": {
-      "created_at": "2023-10-27T10:00:00Z",
-      "author": "dev_user",
-      "version": 1.2
-    }
-  },
-  "employees": [
-    {
-      "id": 1,
-      "name": "Alice Smith",
-      "role": "Lead Developer",
-      "skills": ["Rust", "C++"]
-    },
-    {
-      "id": 2,
-      "name": "Bob Jones",
-      "role": "QA Engineer",
-      "skills": ["Python", "Selenium"]
-    }
-  ],
-  "settings": {
-    "retry_attempts": 5,
-    "timeout_ms": 3000
-  }
-}
\ No newline at end of file
diff --git a/tests_data/scan_directory/test.rs b/tests_data/scan_directory/test.rs
deleted file mode 100644
index 3ea8704d..00000000
--- a/tests_data/scan_directory/test.rs
+++ /dev/null
@@ -1,36 +0,0 @@
-use std::collections::HashMap;
-
-struct User {
-    username: String,
-    email: String,
-    sign_in_count: u64,
-    active: bool,
-}
-
-impl User {
-    fn new(username: String, email: String) -> User {
-        User {
-            username,
-            email,
-            sign_in_count: 1,
-            active: true,
-        }
-    }
-
-    fn deactivate(&mut self) {
-        self.active = false;
-        println!("User {} has been deactivated.", self.username);
-    }
-}
-
-fn main() {
-    let mut users = HashMap::new();
-    let user1 = User::new(String::from("rust_fan"), String::from("rust@example.com"));
-    
-    users.insert("u1", user1);
-
-    match users.get_mut("u1") {
-        Some(u) => u.deactivate(),
-        None => println!("User not found"),
-    }
-}
\ No newline at end of file
diff --git a/tests_data/scan_directory/test.ts b/tests_data/scan_directory/test.ts
deleted file mode 100644
index 5773bc3a..00000000
--- a/tests_data/scan_directory/test.ts
+++ /dev/null
@@ -1,36 +0,0 @@
-interface Task {
-    id: number;
-    title: string;
-    completed: boolean;
-    completedAt?: Date; // Optional property
-}
-
-class TaskManager {
-    private tasks: Task[] = [];
-
-    addTask(title: string): void {
-        const newTask: Task = {
-            id: this.tasks.length + 1,
-            title: title,
-            completed: false
-        };
-        this.tasks.push(newTask);
-    }
-
-    completeTask(id: number): void {
-        const task = this.tasks.find(t => t.id === id);
-        if (task) {
-            task.completed = true;
-            task.completedAt = new Date();
-            console.log(`Task '${task.title}' marked as done.`);
-        }
-    }
-
-    listTasks(): Task[] {
-        return this.tasks;
-    }
-}
-
-const manager = new TaskManager();
-manager.addTask("Learn TypeScript");
-manager.completeTask(1);
\ No newline at end of file

From 3c4b7aa3c4592c62216e0a87e564d1872300b962 Mon Sep 17 00:00:00 2001
From: Anthony Vargas <90121982+Speedrunyourknowledge@users.noreply.github.com>
Date: Tue, 2 Dec 2025 01:18:26 -0500
Subject: [PATCH 13/13] Fix format with ruff

---
 python/src/magika/magika.py | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/python/src/magika/magika.py b/python/src/magika/magika.py
index 0ee892c9..0c616f49 100644
--- a/python/src/magika/magika.py
+++ b/python/src/magika/magika.py
@@ -182,7 +182,11 @@ def scan_directory(
         collected_paths: List[Union[str, os.PathLike]] = []
 
         # Use rglob('*') for recursive scan, glob('*') for single directory
-        glob_pattern = sorted(path_obj.rglob("*")) if recursive_scan else sorted(path_obj.glob("*"))
+        glob_pattern = (
+            sorted(path_obj.rglob("*"))
+            if recursive_scan
+            else sorted(path_obj.glob("*"))
+        )
 
         for item in glob_pattern:
             # We only want files, not sub-directories themselves