troggle-unchained/core/views/uploads.py

import re, os
import subprocess
from pathlib import Path

from django import forms

from django.conf import settings
from django.urls import reverse
from django.db.models import Q
from django.http import HttpResponse, HttpResponseRedirect
from django.shortcuts import render
from django.template import Context, loader
from django.core.files.storage import FileSystemStorage, default_storage

from troggle.parsers.imports import import_caves, import_people, import_surveyscans
from troggle.parsers.imports import import_logbooks, import_QMs, import_drawingsfiles, import_survex
# from databaseReset import reinit_db # don't do this. databaseRest runs code *at import time*
from troggle.core.models.troggle import Expedition, Person, PersonExpedition
from troggle.core.models.caves import LogbookEntry, QM, Cave, PersonTrip
from troggle.core.models.survex import DrawingFile
from .auth import login_required_if_public

'''File upload 'views' 
'''

todo = '''
- Write equivalent photo upload form system, similar to scanupload() but in expofiles/photos/
  Need to validate it as being a valid image file, not a dubious script or hack

- Write equivalent GPX upload form system, similar to scanupload() but in expofiles/gpslogs/
  Need to validate it as being a valid GPX file using an XML parser, not a dubious script or hack
  
- Validate Tunnel & Therion files using an XML parser in dwgupload()

- Validate image files using a magic recogniser in scanupload()

- Enable folder creation in dwguploads or as a separate form

- Register uploaded filenames in the Django db without needing to wait for a reset & bulk file import
'''

class FilesForm(forms.Form):  # not a model-form, just a form-form
    uploadfiles = forms.FileField() 

@login_required_if_public
def scanupload(request, wallet=None):
    '''Upload scanned image files into a wallet on /expofiles
    This does NOT use a Django model linked to a Django form. Just a simple Django form.
    You will find the Django documentation on forms very confusing, This is simpler.
    '''
    filesaved = False
    actual_saved = []
    # print(f'! - FORM scanupload - start {wallet}')
    if wallet is None:
        wallet = "2021#01" # improve this later
    if not re.match('(19|20)\d\d:\d\d', wallet):
        wallet = "2021:01" # improve this later
    
    year = wallet[:4]
    if int(year) < 1977: 
        year = "1977"
    if int(year) > 2050: 
        year = "2050"
    nexty = f'{int(year)+1}'
    prevy = f'{int(year)-1}'

    wnumber = wallet[5:]
    next = f'{int(wnumber)+1:02d}'
    prev = f'{int(wnumber)-1:02d}'
    
    if int(wnumber) == 0:
        prev = f'{int(wnumber):02d}'
    
    context = {'year': year, 'prev': prev, 'next': next, 'prevy': prevy, 'nexty': nexty}
    
    wallet = wallet.replace(':','#')
    dirpath = Path(settings.SURVEY_SCANS, year, wallet)
    
    form = FilesForm()
    
    if request.method == 'POST':
        form = FilesForm(request.POST,request.FILES)
        if form.is_valid():
            f = request.FILES["uploadfiles"]
            multiple = request.FILES.getlist('uploadfiles')
            fs = FileSystemStorage(os.path.join(settings.SURVEY_SCANS, year, wallet))
           
            actual_saved = []
            if multiple:
                for f in multiple:
                    actual_saved.append( fs.save(f.name, content=f) )
                    # print(f'! - FORM scanupload multiple {actual_saved}')
            filesaved = True
    
    files = []
    dirs = []
    # print(f'! - FORM scanupload - start {wallet} {dirpath}')
    try:
        for f in dirpath.iterdir():
            if f.is_dir():
                dirs.append(f.name)
            if f.is_file():
                if f.name != 'contents.json' and f.name != 'walletindex.html':
                    files.append(f.name)
    except FileNotFoundError:
        files.append('(no wallet yet - would be created)')
    if len(files) ==0 :
        files.append('(no image files in wallet)')
    else:
        files = sorted(files)
        
    if dirs:
        dirs = sorted(dirs)

    return render(request, 'scanuploadform.html', 
        {'form': form, 'wallet': wallet, **context, 'files': files, 'dirs': dirs, 'filesaved': filesaved, 'actual_saved': actual_saved})

@login_required_if_public
def dwgupload(request, folder=None, gitdisable='no'):
    '''Upload DRAWING files (tunnel or therion) into the upload folder in :drawings
    AND registers it into the :drawings: git repo.
    
    This does NOT use a Django model linked to a Django form. Just a simple Django form.
    You will find the Django documentation on forms very confusing, This is simpler.
    
    Need to validate it as being a valid GPX file using an XML parser, not a dubious script or hack

    
    We use get_or_create instead of simply creating a new object in case someone uploads the same file
    several times in one session, and expects them to be overwritten in the database. Although
    the actual file will be duplicated in the filesystem with different random name ending.
    '''
    def dwgvalid(name):
        if name in [ '.gitignore', ]:
            return False
        if Path(name).suffix.lower() in ['.xml', '.th', '.th2', '', '.svg', '.txt']:
            return True # dangerous, we should check the actual file binary signature
        return False

    def dwgvaliddisp(name):
        '''OK to display, even if we are not going to allow a new one to be uploaded
        '''
        if name in [ '.gitignore', ]:
            return False
        if Path(name).suffix.lower() in ['.xml', '.th', '.th2', '', '.svg', '.txt', '.jpg', '.jpeg', '.png', '.pdf']:
            return True # dangerous, we should check the actual file binary signature
        return False

    filesaved = False
    actual_saved = []
    refused = []
    doesnotexist = ''
    # print(f'! - FORM dwgupload - start "{folder}" - gitdisable "{gitdisable}"')
    if folder is None:
        folder = "" # improve this later
        dirpath = Path(settings.DRAWINGS_DATA)
        urlfile = '/dwgdataraw'
        urldir  = '/dwgupload'
    else:
        dirpath = Path(settings.DRAWINGS_DATA, folder)
        urlfile = Path('/dwgdataraw/') / folder
        urldir  = Path('/dwgupload/') / folder
    
    form = FilesForm()
    
    if request.method == 'POST':
        form = FilesForm(request.POST,request.FILES)
        if form.is_valid():
            f = request.FILES["uploadfiles"]
            multiple = request.FILES.getlist('uploadfiles')
            fs = FileSystemStorage(os.path.join(settings.DRAWINGS_DATA, folder))
           
            actual_saved = []
            refused = []
            
            # GIT see also core/views/expo.py editexpopage()
            # GIT see also core/models/cave.py writetrogglefile()
            if gitdisable != 'yes': # set in url 'dwguploadnogit/'
                git = settings.GIT
            else:
                git = 'echo'

            if multiple:
                for f in multiple:
                    if dwgvalid(f.name):
                        saved_filename = fs.save(f.name, content=f)
                        actual_saved.append(saved_filename)
                        if gitdisable != 'yes':
                            dr_add = subprocess.run([git, "add", saved_filename], cwd=dirpath, capture_output=True, text=True)
                            if dr_add.returncode != 0:
                                msgdata = 'Ask a nerd to fix this.\n\n' + dr_add.stderr + '\n\n' + dr_add.stdout  + '\n\nreturn code: ' + str(dr_add.returncode)
                                message = f'CANNOT git on server for this file {saved_filename}. Edits saved but not added to git.\n\n' + msgdata
                                return render(request,'errors/generic.html', {'message': message})
                        dwgfile, created = DrawingFile.objects.get_or_create(dwgpath=saved_filename, dwgname=Path(f.name).stem, filesize=f.size)  
                        dwgfile.save()
                    else:
                        refused.append(f.name)
                        print(f'REFUSED {f.name}')
            if actual_saved: # maybe all were refused by the suffix test in dwgvalid()
                filesaved = True
                if len(actual_saved) > 1:
                    dots = "..."
                else:
                    dots = ""
                if gitdisable != 'yes':
                    dr_commit = subprocess.run([git, "commit", "-m", f'Drawings upload - {actual_saved[0]}{dots}'], cwd=dirpath, capture_output=True, text=True)
                    # This produces return code = 1 if it commits OK
                    if dr_commit.returncode != 0:
                        msgdata = 'Ask a nerd to fix this.\n\n' + dr_commit.stderr + '\n\n' + dr_commit.stdout  + '\n\nreturn code: ' + str(dr_commit.returncode)
                        message = f'Error code with git on server for this {actual_saved[0]}{dots}. Edits saved, added to git, but NOT committed.\n\n' + msgdata
                        return render(request,'errors/generic.html', {'message': message})

    files = []
    dirs = []
    # print(f'! - FORM dwgupload - start {folder} \n"{dirpath}" \n"{dirpath.parent}" \n"{dirpath.exists()}"')
    try:
        for f in dirpath.iterdir():
            if f.is_dir():
                if f.name not in ['.git' ]:
                    dirs.append(f.name)
                    continue
            if f.is_file():
                if dwgvaliddisp(f.name): 
                    files.append(f.name)
                    continue
    except FileNotFoundError:
        doesnotexist = True
    if files:
        files = sorted(files)
        
    if dirs:
        dirs = sorted(dirs)

    return render(request, 'dwguploadform.html', 
        {'form': form, 'doesnotexist': doesnotexist, 'urlfile': urlfile, 'urldir': urldir,'folder': folder, 'files': files, 'dirs': dirs, 'filesaved': filesaved, 'actual_saved': actual_saved, 'refused': refused})