pretty-breaking-changes-markdown.py

#! /usr/bin/python3

import git
import mistune
import os
# import re

repo_path = "/home/yo/src/liferay-portal-ee"

start_hash = "f63d698232b7b620536bb32f854286b132fcc07a"
end_hash = "4a2e75e4e6fd4c225db62f303ccc33cdd1782ba4"

amendments_file_path = repo_path + "/readme/BREAKING_CHANGES_AMENDMENTS.markdown"

template_path = "/home/yo/projects/pretty-breaking-changes"

breaking_change_report_keyword = "# breaking"

markdown = mistune.create_markdown(renderer='ast')

def get_end_of_block_using_pattern(lines, start_index, end_pattern):
    if start_index >= len(lines):
        return start_index
    
    end_pattern_lowercase = end_pattern.lower()
    
    end_of_block_index = start_index
    
    while end_of_block_index < len(lines):
        if not lines[end_of_block_index].lower().startswith(end_pattern_lowercase):
            end_of_block_index += 1
        else:
            break
    
    # print(end_of_block_index)
    return end_of_block_index
         
def get_end_of_block(lines, start_index):
    next_line_index = get_end_of_block_using_pattern(lines, start_index, '----')
    
    if next_line_index == len(lines):
        next_line_index = get_end_of_block_using_pattern(lines, start_index, breaking_change_report_keyword)
        
    return next_line_index
            
         
def is_empty_block(lines):
    result = True
    
    for line in lines:
        if line.strip():
            return False
    
    return result

def dissect_commit_message(raw_commit_message):
    breaking_changes_info = []
    
    lines = raw_commit_message.split('\n')
    
    raw_jira_info_block_line_index = get_end_of_block_using_pattern(lines, 1, breaking_change_report_keyword)
    
    i = 0
    while i < raw_jira_info_block_line_index:
        if lines[i] != "":
            jira_ticket, _, jira_ticket_title = lines[i].partition(' ')
            break
        i += 1
    
    block_start_line_index = raw_jira_info_block_line_index
    
    while block_start_line_index < len(lines):
        # print(str(block_start_line_index) + " " + str(len(lines)))
        next_line_index = get_end_of_block(lines, block_start_line_index)
        
        # print(next_line_index)
        if next_line_index - block_start_line_index > 1 and not is_empty_block(lines[block_start_line_index:next_line_index]):
            info = None
            
            if lines[next_line_index - 1] == '----':
                block_end_line_index = next_line_index - 1
            else:
                block_end_line_index = next_line_index
            
            try:
                info = extract_breaking_change_info(lines[block_start_line_index:block_end_line_index])
            except Exception as e:
                print("unprocessable")
                print(e)
                print(lines)
                print("Range considered: " + str(block_start_line_index) + ":" + str(block_end_line_index))
                print()
            
            if info:
                breaking_changes_info.append(info)
                breaking_changes_info[-1]['jira_ticket'] = jira_ticket
                breaking_changes_info[-1]['jira_ticket_title'] = jira_ticket_title.rstrip('\r\n')
                
        block_start_line_index = next_line_index + 1
        
    return breaking_changes_info
    

def extract_breaking_change_info(lines):
    # print(lines)
    breaking_change_info = {}
    
    what_line_index = get_end_of_block_using_pattern(lines, 0, "## What")
    
    if what_line_index >= len(lines): 
        raise LookupError('"## What" not found')
    
    breaking_change_info['affected_file_path'] = get_affected_file_path(lines[what_line_index])
    
    why_line_index = get_end_of_block_using_pattern(lines, what_line_index + 1, "## Why")
    
    if why_line_index >= len(lines):
        raise LookupError('"## Why" not found')
    
    breaking_change_info['what_info'] = "\n".join(lines[what_line_index + 1:why_line_index]).rstrip('\r\n')
    
    alternatives_line_index = get_end_of_block_using_pattern(lines, why_line_index + 1, "## Alternatives")
    
    breaking_change_info['why_info'] = "\n".join(lines[why_line_index + 1: alternatives_line_index]).rstrip('\r\n')
    
    if len(lines) > alternatives_line_index:
        breaking_change_info['alternatives'] = "\n".join(lines[alternatives_line_index + 1: len(lines)]).rstrip('\r\n')
    
    return breaking_change_info

def decorate_breaking_change_info(breaking_change_info, decoration):
    return [item|decoration for item in breaking_change_info]

def get_affected_file_path(raw_what_header):
    _, _, affected_file_path = raw_what_header.partition('## What ')
    
    if affected_file_path == "":
        _, _, affected_file_path = raw_what_header.partition('## what ')
        
    return affected_file_path

def get_first_level_path(file_path):
    first_level_path, _, remaining = file_path.partition('/')
    
    if len(first_level_path) == 0:
        first_level_path, _, remaining = remaining.partition('/')
        
    if len(remaining) == 0:
        first_level_path = 'other'
        
    return first_level_path

liferay_portal_ee_repo = git.Repo(repo_path)

print("Retrieving git info ...")

of_interest = liferay_portal_ee_repo.git.log(start_hash + ".." + end_hash, "--grep", "# breaking", "--pretty=format:%H")
# of_interest = liferay_portal_ee_repo.git.log("--grep", "breaking_change_report", "--pretty=format:%H")

print("Processing git info ...")

individual_commit_hashes = of_interest.split('\n')

breaking_changes_info = {h:decorate_breaking_change_info(result, {'committed_date': liferay_portal_ee_repo.commit(h).committed_date}) for h in individual_commit_hashes if (result := dissect_commit_message(liferay_portal_ee_repo.commit(h).message))}

amendments = ""

with open(amendments_file_path) as f:
    amendments = f.read()

parsed = markdown.parse(amendments)
interesting_indexes = [(i, type_of) for i in range(len(parsed)) if (type_of := parsed[i]['type']) == 'heading' or type_of == 'block_code']

i = 0
while i <= len(interesting_indexes) - 2:
    if interesting_indexes[i][1] == 'heading' and interesting_indexes[i + 1][1] == 'block_code':
        hash = parsed[interesting_indexes[i][0]]['children'][0]['text']
        
        if hash in breaking_changes_info or (liferay_portal_ee_repo.is_ancestor(start_hash, hash) and not liferay_portal_ee_repo.is_ancestor(end_hash, hash)):
            amended_message = parsed[interesting_indexes[i + 1][0]]['text']
            breaking_changes_info[hash] = decorate_breaking_change_info(dissect_commit_message(amended_message), {'committed_date': liferay_portal_ee_repo.commit(hash).committed_date})
            
            
            if len(breaking_changes_info[hash]) > 0:
                print("Amending: " + str(breaking_changes_info[hash][0]['jira_ticket']))
            else:
                print("Error processing amendment message " + hash)
                print(amended_message)
                print()
        
        i += 2
    else:
        i += 1
    

affected_file_paths_and_hashes = {}

for hash in breaking_changes_info:
    # print(hash)
    # print(breaking_changes_info[hash]['affected_file_path'])
    
    for change in breaking_changes_info[hash]:
        affected_file_path = change['affected_file_path']
        # print(affected_file_path)
        first_level_path = get_first_level_path(affected_file_path)
            
        if not first_level_path in affected_file_paths_and_hashes:
            affected_file_paths_and_hashes[first_level_path] = {}
        
        if affected_file_path in affected_file_paths_and_hashes[first_level_path]:
            affected_file_paths_and_hashes[first_level_path][affected_file_path].append(change)
        else:
            affected_file_paths_and_hashes[first_level_path][affected_file_path] = [change]

print("Generating output ...")

entire_output = ""

for first_level_path in affected_file_paths_and_hashes:
    for affected_file_path in affected_file_paths_and_hashes[first_level_path]:
        file_name = os.path.basename(affected_file_path)
        this_block = f'''
# {affected_file_path}
          
{file_name} `{affected_file_path}`
        '''
        
        for change in affected_file_paths_and_hashes[first_level_path][affected_file_path]:
            this_block += '''
* __Date:__ {committed_date}
* __Ticket:__ [{jira_ticket}](https://liferay.atlassian.net/browse/{jira_ticket})
* __What changed:__ {what_info}
* __Reason:__ {why_info}
            '''.format(**change)
            
            if 'alternatives' in change:
                this_block += '''
* __Alternatives:__ {alternatives}
                '''.format(**change)
                
            this_block += '''
&nbsp;
            '''
    
        entire_output += this_block

entire_output_fh = open('report.md', 'w')

entire_output_fh.write(entire_output)

entire_output_fh.close()