make_diff.py

#!/usr/bin/env python3
# AGPL license, if used in Netbox+its forks Apache-2.0
import json
import html


def regular_line_ending(s: str):
    if s.endswith(","):
        return True
    return not (s.endswith("{") or s.endswith("["))


def begin_of_complex(s: str):
    return s.endswith("[") or s.endswith("{")


def has_indent(s: str, indent: str):
    return not s.removeprefix(indent).startswith(" ")


def make_diff(old: str, new: str):
    old_lines = old.splitlines(False)
    new_lines = new.splitlines(False)
    old_list = []
    new_list = []
    old_idx = 0
    new_idx = 0
    while True:
        if old_idx == len(old_lines) and new_idx == len(new_lines):
            break
        if old_idx == len(old_lines):
            while new_idx != len(new_lines):
                new_list.append((new_lines[new_idx], True))
                old_list.append(("", False))
                new_idx += 1
            break
        if new_idx == len(new_lines):
            while old_idx != len(old_lines):
                old_list.append((old_list[old_idx], True))
                new_list.append(("", False))
                old_idx += 1
            break
        old_s = old_lines[old_idx]
        old_idx += 1
        new_s = new_lines[new_idx]
        new_idx += 1
        if "{" == old_s.strip() and "{" == new_s.strip():
            old_list.append(("{", False))
            new_list.append(("{", False))
            continue
        if "}" == old_s.strip() and "}" == new_s.strip():
            old_list.append(("}", False))
            new_list.append(("}", False))
            continue
        # Handle:
        #  "foo": null    "foo": [
        #                     "bar",
        #                     "baz",
        #                 "]"
        if regular_line_ending(old_s) and begin_of_complex(new_s):
            indent_of_new = new_s.replace(new_s.strip(), "")
            new_list.append((new_s, True))
            old_list.append((old_s, True))
            while not has_indent(new_lines[new_idx], indent_of_new):
                old_list.append(("", True))
                new_list.append((new_lines[new_idx], True))
                new_idx += 1
            old_list.append(("", True))
            new_list.append((new_lines[new_idx], True))
            new_idx += 1
            continue
        # Handle:
        # "foo": {         "foo": null
        #   "foo": "bar"
        # }
        if begin_of_complex(old_s) and regular_line_ending(new_s):
            indent_of_old = old_s.replace(old_s.strip(), "")
            old_list.append((old_s, True))
            new_list.append((new_s, True))
            while not has_indent(old_lines[old_idx], indent_of_old):
                new_list.append(("", True))
                old_list.append((old_lines[old_idx], True))
                old_idx += 1
            new_list.append(("", True))
            old_list.append((old_lines[old_idx], True))
            old_idx += 1
            continue
        if begin_of_complex(old_s) and begin_of_complex(new_s):
            indent_of_old = old_s.replace(old_s.strip(), "")
            indent_of_new = new_s.replace(new_s.strip(), "")
            old_list.append((old_s, False))
            new_list.append((new_s, False))
            print(">>", new_list[-1], new_s)
            old_tmp = []
            new_tmp = []
            while not has_indent(old_lines[old_idx], indent_of_old):
                old_tmp.append(old_lines[old_idx])
                old_idx += 1
            while not has_indent(new_lines[new_idx], indent_of_new):
                new_tmp.append(new_lines[new_idx])
                new_idx += 1
            a, b = make_diff("\n".join(old_tmp), "\n".join(new_tmp))
            assert len(a) == len(b)
            old_list += a
            new_list += b
            old_list.append((old_lines[old_idx], False))
            old_idx += 1
            new_list.append((new_lines[new_idx], False))
            new_idx += 1
            continue
        if old_s == new_s:
            old_list.append((old_s, False))
            new_list.append((new_s, False))
            continue
        # Handle:
        # "foo": 1(,)  "foo": 2(,)
        if regular_line_ending(old_s) and regular_line_ending(new_s):
            old_list.append((old_s, True))
            new_list.append((new_s, True))
            continue
    return old_list, new_list


def format_html(tup, color):
    if tup[0] == "":
        return f"<pre>{html.escape('<<EMPTY>>')}</pre>"
    if not tup[1]:
        return f"<pre>{html.escape(tup[0])}</pre>"
    return f'<pre style="background-color: {color}">{html.escape(tup[0])}</pre>'


old = {
    "a": 5,
    "b": 2,
    "c": [],
    "d": None,
    "e": {"foo": "baz"},
    "f": [1],
    "g": [1, 2, 3],
    "h": {"a": "b", "c": "d", "q": "r"},
}
new = {
    "a": 5,
    "b": None,
    "c": None,
    "d": [1],
    "e": None,
    "f": [1, 2],
    "g": [],
    "h": {"a": "c", "c": "d", "f": "g", "q": "r"},
}
print(json.dumps(old, indent=2, sort_keys=True))
print(json.dumps(new, indent=2, sort_keys=True))
a, b = make_diff(
    json.dumps(old, indent=2, sort_keys=True), json.dumps(new, indent=2, sort_keys=True)
)
print(a)
print(b)
assert len(a) == len(b)
a_html = "\n".join([format_html(x, "red") for x in a])
b_html = "\n".join([format_html(x, "green") for x in b])
with open("diff.html", "w", encoding="utf-8") as filep:
    filep.write(
        f"""
            <!DOCTYPE html>
            <html>
            <head>
            <meta name=\"viewport\" content=\"width=device-width, initial-scale=1\">
            <style>
            * {{
            box-sizing: border-box;
            }}

            /* Create two equal columns that floats next to each other */
            .column {{
            float: left;
            width: 50%;
            padding: 10px;
            height: 100%;
            }}
            .row:after {{
            content: \"\";
            display: table;
            clear: both;
            }}
            </style>
            </head>
            <body>

            <h2>Two Equal Columns</h2>

            <div class="row">
            <div class="column" style="background-color:#aaa;">
                <h2>Column 1</h2>
            {a_html}
            </div>
            <div class="column" style="background-color:#bbb;">
                <h2>Column 2</h2>
            {b_html}
            </div>
            </div>

            </body>
            </html>

    """
    )