#!/usr/bin/python

import os
import re

#this script removes md files which contain errors to get a successful build. has to be done stage by stage because some errors only occur after other ones are fixed

docusaurusPath = '/docusaurus/' + os.environ['WEBSITE_NAME'] + '/'
stdout = docusaurusPath + 'errorlog.txt'
stderr = docusaurusPath + 'linkslog.txt'
removedDataPath = docusaurusPath + 'docs/RemovedData.md'
brokenLinksPath = docusaurusPath + 'docs/BrokenLinks.md'

def getProblematicMds(log):
    problematicMds = {}
    def extractData(error):
        queryErrorPath = "REGEXQUERYWHICHDISCLOSESCOMPANYSENSITIVEDATA"
        querySyntaxErrorPath = 'SyntaxError:.*\.md'
        resultErrorPath = re.search(queryErrorPath, error)
        resultSyntaxErrorPath = re.search(querySyntaxErrorPath, error)
        if resultErrorPath != None:
            problematicMds[docusaurusPath + resultErrorPath.group()[4:]] = error
        elif resultSyntaxErrorPath != None:
            problematicMds[resultSyntaxErrorPath.group()[13:]] = error    

    queryErr = "(Error|SyntaxError)(.*?)((?=Error)|(?=SyntaxError))"  
    errors = list(map(lambda tuple: tuple[0] + tuple[1], re.findall(queryErr, log, flags = re.S | re.M)))
    if len(errors) != 0:
        for i, error in enumerate(errors):
            extractData(error)
            #for the last error, just take the index of the error before and assume this is the error message (queryErr regex doesnt catch last error)
            if i == len(errors) - 1:
                lastError = log[log.rindex(error) + len(error):]
                extractData(lastError)
    else:
        extractData(log)

    return problematicMds

def deleteProblematicMds(foundMds):
    for md in foundMds:
        if md[0:22] != '/docusaurus/MyWebsite/':
            md = '/docusaurus/MyWebsite' + md
        if md[-3:] != '.md':
            md += '.md'
        if os.path.exists(md):
            print('deleted md: ' + md + '  ')
            os.remove(md)
        else:
            print('cannot delet md: ' + md + ' ')

def createAppendMdOfRemoved(foundMds, header):
    append_write = 'a' if os.path.exists(removedDataPath) else 'w'
    f = open(removedDataPath, append_write)
    if append_write == 'w':
        f.write('# Removed files/ directories \n')    
    f.write(header + '\n')
    if isinstance(foundMds, dict):
        for path, errormsg in foundMds.items():
            f.write('<details>\n<summary>' + path + '</summary>\n\n```\n' + errormsg + '\n```\n\n</details>\n\n')
    else:
        for md in foundMds:
            f.write('* ' + md[md.rfind('/'):] + '  \n')
    f.close()

def createMdOfBrokenLinks(log):
    queryWarningMessage = 'warn(.*?)in(.*?)\.md(.*?)\.md'
    queryBrokenLink = '\:.*?\"'
    queryContainedFile = '\/docusaurus\/.*\''
    brokenLinks = re.findall(queryWarningMessage, log)
    if len(brokenLinks) != 0:
        f = open(brokenLinksPath, "w")
        f.write("## Broken links\n|File path|Broken link|\n|--|--|\n")
        for link in brokenLinks:
            containedFile = re.search(queryContainedFile, str(link)).group()[40:-1]
            brokenLink = re.search(queryBrokenLink, str(link)).group()[3:-1]
            f.write('|' + containedFile + '|' + brokenLink + '|\n')
        f.close()

def mdFormatProblem(log):
    #dont always write the full log, filter it for the specific files that have problems
    print('Problems in MD files.. deleting')
    problematicMds = getProblematicMds(log)
    deleteProblematicMds(problematicMds.keys())
    createAppendMdOfRemoved(problematicMds, '## Markdown format problems')
    print('starting build process again..')
    runSaurus()

def removeFilesByQuery(log, msg, query):
    print('Having ' + msg)
    problematicMds = list(re.findall(query, log, flags = re.S | re.M))
    deleteProblematicMds(problematicMds)
    createAppendMdOfRemoved(problematicMds, '## ' + msg)
    print('starting build process again')
    runSaurus()

def runSaurus():
    result = os.system( 'yarn run build 2> ' + stderr + ' > ' + stdout)
    outlog = open(stdout, "r").read()
    errlog = open(stderr, "r").read()
    createMdOfBrokenLinks(errlog)
    print('Return value from building: ' + str(result))
    #error 2 = manual abortion, useful for local debugging to intercept an infinite loop    
    if result == 2:
        exit(0)
    elif result != 0:
        if 'Client: Compiled successfully' in outlog:            
            if 'could not render static page with path' in errlog:
                #case: all mds SEEM to be formatted correctly according to docusaurus, still not running through bc of 
                # 'Docusaurus Node/SSR could not render static page with path ...' errors
                removeFilesByQuery(errlog, 'forbidden file content problems', 'could not render static page with path "(.*?)(?=\")')
            else:
                print('Successful build!')
                #still throws an error but compiles and produces the static html files successfully. is smth missing?
        else:
            if 'Error while parsing Markdown frontmatter' in errlog:
                #another kind of error which is caused by some yaml translation
                removeFilesByQuery(errlog, 'errors while parsing Markdown frontmatter', 'doc metadatas for doc at path \"(.*?)(?=\" in version)')
            else:
                #some mds have an incorrect format
                mdFormatProblem(outlog)
    else:
        print('Successful build!')

runSaurus()