Refactoring ssg.py to improve code maintainability

* Changed a word in a comment * Simplified markdown header parsing * Extracted SSGParser class * Extracted SSGUtil class
a-rokay · Oct 14, 2021 · 0ffa06c · 0ffa06c
1 parent fb0fd72
commit 0ffa06c
Show file tree

Hide file tree

Showing 6 changed files with 1,319 additions and 151 deletions.
diff --git a/.gitignore b/.gitignore
@@ -1 +1,2 @@
-ssg/dist/
+ssg/dist/
+__pycache__
diff --git a/ssg/SSGParser.py b/ssg/SSGParser.py
@@ -0,0 +1,91 @@
+import re
+import SSGUtil
+
+html_skeleton = """<!doctype html>
+<html lang="{lang}">
+<head>
+    <meta charset="utf-8">
+    <title>{title}</title>
+    <meta name="viewport" content="width=device-width, initial-scale=1">{stylesheet}
+</head>
+<body>
+{content}
+</body>
+</html>
+"""
+
+# Returns None if no title is found
+def get_txt_title(file_location):
+    title = None
+    first_3_lines = []
+    i = 0
+
+    # Read file line by line, up to line 3
+    with open(file_location, "r", encoding="utf8") as input_file:
+        while(line := input_file.readline()):
+
+            if(i == 3):
+                break
+
+            cleaned_line = line.rstrip()
+            first_3_lines.append(cleaned_line)
+            i += 1
+
+    # Check if title criteria is met
+    if(len(first_3_lines) == 3):
+        if(first_3_lines[0] and not first_3_lines[1] and not first_3_lines[2]):
+            title = first_3_lines[0]
+
+    return title
+
+# Returns html content, including title, if set
+def generate_content(file_location, title):
+    if(not SSGUtil.is_file_accepted(file_location)):
+        return
+
+    titled_format = "<h1>{}</h1>\n\n\n{}"
+    content = ""
+
+    with open(file_location, "r", encoding="utf8") as input_file:
+        content = input_file.read()
+
+    # Skip the first 3 lines if there's a title
+    if(title):
+        content = content.split("\n", 3)[3]
+
+    content = "<p>" + content
+    content = content.replace("\n\n", "</p>\n\n<p>")
+    content = content + "\n</p>"
+
+    if file_location.endswith(".md"):
+        content = process_markdown(content)
+
+    if(title):
+        content = titled_format.format(title, content)
+
+    return content
+
+def process_markdown(content):
+    # Process bold markdown
+    content = re.sub('\*\*([^\s\*.].*?)\*\*|__([^\s_.].*?)__', r'<strong>\1</strong>', content, flags=re.DOTALL)
+
+    # Process italic markdown
+    content = re.sub('\*([^\s\*.].*?)\*|_([^\s\_.].*?)_', r'<em>\1\2</em>', content, flags=re.DOTALL)
+
+    # Process single backtick markdown
+    content = re.sub('`([^\r\n\`].*?)`', r'<code>\1</code>', content, flags=re.DOTALL)
+
+    # Process header markdown
+    headerTag = lambda s: '<h{size}>{content}</h{size}>\n'.format(size=s.group(2).count('#'), content=s.group(3))
+    content = re.sub(r'(|(?<!\n)\n|<p>)(#{1,5})\s(.*)(<\/p>|(?<!<\/p>)\n|$)', headerTag, content)
+
+    return content
+
+# Inserts title, stylesheet, and content to html_skeleton, returns the result
+def generate_html(lang, file_name, title, stylesheet, content):
+    ss_tag = ""
+
+    if(stylesheet):
+        ss_tag = '\n\t<link rel="stylesheet" href="{}">'.format(stylesheet)
+
+    return html_skeleton.format(lang=lang, title=title if title else file_name, stylesheet=ss_tag, content=content)
diff --git a/ssg/SSGUtil.py b/ssg/SSGUtil.py
@@ -0,0 +1,51 @@
+import json
+import os
+
+OUTPUT_FOLDER = "dist"
+ACCEPTED_FILE_TYPES = [".txt", ".md"]
+
+# Returns a list of files with acceptable file types by filtering all files in folder_name
+def get_accepted_files(folder_name):
+    all_files = os.listdir(folder_name)
+    filtered_files = filter(lambda f: is_file_accepted(f), all_files)
+    return list(filtered_files)
+
+def is_file_accepted(filename):
+    for type in ACCEPTED_FILE_TYPES:
+        if filename.endswith(type):
+            return True
+    return False
+
+def output_to_file(file_name, html):
+    # Create output folder if it doesn't exist
+    if(not os.path.isdir(OUTPUT_FOLDER)):
+        os.mkdir(OUTPUT_FOLDER)
+
+    file_location = OUTPUT_FOLDER + "/" + file_name.replace(file_name[file_name.rfind("."):], ".html")
+    with open(file_location, "w", encoding="utf8") as output_file:
+        output_file.write(html)
+
+def get_config(config, input, lang, stylesheet):
+    # Check to see if config file exists.
+    if config:
+        with open(config) as f:
+            try:
+                data = json.load(f)
+                if len(data) == 0:
+                    print("\nConfig File is empty!\n")
+                    exit(1)
+            except (json.decoder.JSONDecodeError) as err:
+                print("\nError parsing Config File: {0}\n".format(err))
+                exit(1)
+        # For each command from JSON config file, set local variables
+        for i in data:
+            if i == "input" or i == "i":
+                input = data[i]
+            elif i == "stylesheet" or i == "s":
+                stylesheet = data[i]
+            elif i == "lang" or i == "l":
+                lang = data[i]
+        if input == None:
+            print("No input file specified")
+            exit(1)
+    return input, lang, stylesheet