1
0
Fork 0
flightgear/scripts/python/TerraSync/terrasync/main.py
Florent Rougon 13f943b4a1 terrasync.py: rename DirIndex attributes and remove accessors
In Python, common usage is not to define accessors, but to directly use
class or instance attributes (especially when the associated data is
constant after instance creation). If it later happens that a given
attribute needs getter or setter logic, this can always be done via the
@property decorator, and doesn't affect calling code at all. See for
instance:

  https://docs.python.org/3/library/functions.html#property
  https://mail.python.org/pipermail/tutor/2012-December/thread.html#92990

Apply this to the DirIndex class and rename the following attributes for
better readability: f -> files, d -> directories, t -> tarballs.
2020-10-02 16:38:08 +02:00

721 lines
28 KiB
Python
Executable file

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
#
# main.py --- Main module for terrasync.py
#
# Copyright (C) 2016 Torsten Dreyer
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License as
# published by the Free Software Foundation; either version 2 of the
# License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
# General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
import argparse
import enum
import hashlib
import os
import re
import shutil
import sys
import time
import urllib
from urllib.parse import urlparse, urljoin
from http.client import HTTPConnection, _CS_IDLE, HTTPException
from os import listdir
from os.path import isfile, isdir, join
from . import dirindex
from .exceptions import UserError, NetworkError, RepoDataError, \
InvalidDirIndexFile
from .virtual_path import VirtualPath
PROGNAME = os.path.basename(sys.argv[0])
class ExitStatus(enum.Enum):
SUCCESS = 0
# The program exit status is 1 when an exception isn't caught.
ERROR = 1
CHECK_MODE_FOUND_MISMATCH = 2
# *****************************************************************************
# * Utility functions *
# *****************************************************************************
# If a path matches this regexp, we really don't want to delete it recursively
# (“cre” stands for “compiled regexp”).
_removeDirectoryTree_dangerous_cre = re.compile(
r"""^(/ (home (/ [^/]*)? )? /* | # for Unix-like systems
[a-zA-Z]: [\/]* # for Windows
)$""", re.VERBOSE)
def removeDirectoryTree(base, whatToRemove):
"""Recursively remove directory 'whatToRemove', with safety checks.
This function ensures that 'whatToRemove' does not resolve to a
directory such as /, /home, /home/foobar, C:\, d:\, etc. It is also
an error if 'whatToRemove' does not literally start with the value
of 'base' (IOW, this function refuses to erase anything that is not
under 'base').
'whatToRemove' is *not* interpreted relatively to 'base' (this would
be doable, just a different API).
"""
assert os.path.isdir(base), "Not a directory: {!r}".format(base)
assert (base and
whatToRemove.startswith(base) and
whatToRemove[len(base):].startswith(os.sep)), \
"Unexpected base path for removeDirectoryTree(): {!r}".format(base)
absPath = os.path.abspath(whatToRemove)
if _removeDirectoryTree_dangerous_cre.match(absPath):
raise UserError("in order to protect your data, refusing to "
"recursively delete '{}'".format(absPath))
else:
shutil.rmtree(absPath)
def computeHash(fileLike):
hash = hashlib.sha1()
for chunk in iter(lambda: fileLike.read(4096), b""):
hash.update(chunk)
return hash.hexdigest()
def hashForFile(fname):
with open(fname, "rb") as f:
return computeHash(f)
# *****************************************************************************
# * Network-related classes *
# *****************************************************************************
class HTTPGetCallback:
def __init__(self, src, callback):
"""Initialize an HTTPGetCallback instance.
src -- a VirtualPath instance (corresponding to the path on
the server for which a GET request is to be issued)
callback -- a function taking two parameters: the URL (string)
and an http.client.HTTPResponse instance. When
invoked, the callback return value will be returned
by HTTPGetter.get().
"""
self.callback = callback
self.src = src
class HTTPGetter:
def __init__(self, baseUrl, maxPending=10):
self.baseUrl = baseUrl
self.parsedBaseUrl = urlparse(baseUrl)
self.maxPending = maxPending
self.requests = []
self.pendingRequests = []
self.httpConnection = HTTPConnection(self.parsedBaseUrl.netloc)
self.httpRequestHeaders = headers = {'Host':self.parsedBaseUrl.netloc,'Content-Length':0,'Connection':'Keep-Alive','User-Agent':'FlightGear terrasync.py'}
def assemblePath(self, httpGetCallback):
"""Return the path-on-server for the file to download.
Example: '/scenery/Airports/N/E/4/.dirindex'
"""
assert not self.parsedBaseUrl.path.endswith('/'), \
repr(self.parsedBaseUrl)
return self.parsedBaseUrl.path + str(httpGetCallback.src)
def assembleUrl(self, httpGetCallback):
"""Return the URL of the file to download."""
baseUrl = self.parsedBaseUrl.geturl()
assert not baseUrl.endswith('/'), repr(baseUrl)
return urljoin(baseUrl + '/', httpGetCallback.src.asRelative())
def doGet(self, httpGetCallback):
conn = self.httpConnection
pathOnServer = self.assemblePath(httpGetCallback)
self.httpConnection.request("GET", pathOnServer, None,
self.httpRequestHeaders)
httpResponse = self.httpConnection.getresponse()
# 'httpResponse' is an http.client.HTTPResponse instance
return httpGetCallback.callback(self.assembleUrl(httpGetCallback),
httpResponse)
def get(self, httpGetCallback):
nbRetries = nbRetriesLeft = 5
while True:
try:
return self.doGet(httpGetCallback)
except HTTPException as exc:
if nbRetriesLeft == 0:
raise NetworkError(
"after {nbRetries} retries for URL {url}: {errMsg}"
.format(nbRetries=nbRetries,
url=self.assembleUrl(httpGetCallback),
errMsg=exc)) from exc
# Try to reconnect
self.httpConnection.close()
time.sleep(1)
self.httpConnection.connect()
nbRetriesLeft -= 1
class HTTPDownloadRequest(HTTPGetCallback):
def __init__(self, terrasync, src, dst, callback = None ):
"""Initialize an HTTPDownloadRequest instance.
terrasync -- a TerraSync instance
src -- a VirtualPath instance (corresponding to the path
on the server for which a GET request is to be
issued)
dst -- file path (or whatever open() accepts) where the
downloaded data is to be stored
callback -- a function that will be called if the download is
successful, or None if no such callback is desired.
The function must take one parameter: when invoked,
it will be passed this HTTPDownloadRequest
instance. Its return value is ignored.
"""
super().__init__(src, self.callback)
self.terrasync = terrasync
self.dst = dst
self.mycallback = callback
# 'httpResponse' is an http.client.HTTPResponse instance
def callback(self, url, httpResponse):
# I suspect this doesn't handle HTTP redirects and things like that. As
# mentioned at <https://docs.python.org/3/library/http.client.html>,
# http.client is a low-level interface that should normally not be used
# directly!
if httpResponse.status != 200:
raise NetworkError("HTTP callback got status {status} for URL {url}"
.format(status=httpResponse.status, url=url))
try:
with open(self.dst, 'wb') as f:
f.write(httpResponse.read())
except HTTPException as exc:
raise NetworkError("for URL {url}: {error}"
.format(url=url, error=exc)) from exc
if self.mycallback is not None:
self.mycallback(self)
class HTTPSocketRequest(HTTPGetCallback):
"""HTTPGetCallback class whose callback returns a file-like object.
The file-like object returned by the callback, and thus by
HTTPGetter.get(), is a socket or similar. This allows one to read
the data obtained from the network without necessarily storing it
to a file.
"""
def __init__(self, src):
"""Initialize an HTTPSocketRequest object.
src -- VirtualPath instance for the resource on the server
(presumably a file)
"""
HTTPGetCallback.__init__(self, src, self.callback)
def callback(self, url, httpResponse):
# Same comment as for HTTPDownloadRequest.callback()
if httpResponse.status != 200:
raise NetworkError("HTTP callback got status {status} for URL {url}"
.format(status=httpResponse.status, url=url))
return httpResponse
#################################################################################################################################
class Coordinate:
def __init__(self, lat, lon):
self.lat = lat
self.lon = lon
class DownloadBoundaries:
def __init__(self, top, left, bottom, right):
if top < bottom:
raise ValueError("top cannot be less than bottom")
if right < left:
# right may be less than left when wrapping across the antimeridian
if not (left >= 0 and right < 0):
raise ValueError("right cannot be less than left")
if top > 90 or bottom < -90:
raise ValueError("top and bottom must be a valid latitude")
if left < -180 or right >= 180:
raise ValueError("left and right must be a valid longitude")
self.top = top
self.left = left
self.bottom = bottom
self.right = right
def is_coordinate_inside_boundaries(self, coordinate, isOuterBucket):
bigTileBottom = coordinate.lat
bigTileTop = bigTileBottom + (10 if isOuterBucket else 1)
bigTileLeft = coordinate.lon
bigTileRight = bigTileLeft + (10 if isOuterBucket else 1)
# if the two regions do not overlap then we are done
if bigTileTop <= self.bottom or bigTileBottom > self.top:
return False
if bigTileRight <= self.left or bigTileLeft > self.right:
# check for spanning across the antimeridian
if self.left >= 0 and self.right < 0:
# determine which side we are on and check of region overlap
if bigTileLeft >= 0:
if bigTileRight <= self.left:
return False
elif bigTileLeft > self.right:
return False
else:
return False
# at least a partial overlap exists, so more processing will be needed
return True
def parse_terrasync_coordinate(coordinate):
matches = re.match("(w|e)(\d{3})(n|s)(\d{2})", coordinate)
if not matches:
return None
lon = int(matches.group(2))
if matches.group(1) == "w":
lon *= -1
lat = int(matches.group(4))
if matches.group(3) == "s":
lat *= -1
return Coordinate(lat, lon)
class Report:
"""Gather and format data about the state of a TerraSync mirror."""
def __init__(self, targetDir):
self.targetDir = targetDir
self.dirsWithMissingIndex = set()
self.dirsWithMismatchingDirIndexHash = set()
self.missingFiles = set()
self.filesWithMismatchingHash = set()
self.dirsSkippedDueToBoundaries = set()
self.orphanFiles = set()
self.orphanDirs = set()
def addMissingDirIndex(self, directoryVirtualPath):
self.dirsWithMissingIndex.add(directoryVirtualPath)
def addDirIndexWithMismatchingHash(self, directoryVirtualPath):
self.dirsWithMismatchingDirIndexHash.add(directoryVirtualPath)
def addMissingFile(self, virtualPath):
self.missingFiles.add(virtualPath)
def addFileWithMismatchingHash(self, virtualPath):
self.filesWithMismatchingHash.add(virtualPath)
def addSkippedDueToBoundaries(self, virtualPath):
self.dirsSkippedDueToBoundaries.add(virtualPath)
def addOrphanFile(self, virtualPath):
self.orphanFiles.add(virtualPath)
def addOrphanDir(self, virtualPath):
self.orphanDirs.add(virtualPath)
def summaryString(self):
reportElements = [
("Directories with missing index", self.dirsWithMissingIndex),
("Directories whose .dirindex file had a mismatching hash",
self.dirsWithMismatchingDirIndexHash),
("Missing files", self.missingFiles),
("Files with a mismatching hash", self.filesWithMismatchingHash),
("Directories skipped because of the specified boundaries",
self.dirsSkippedDueToBoundaries),
("Orphan files", self.orphanFiles),
("Orphan directories", self.orphanDirs)]
l = []
for heading, setOfFilesOrDirs in reportElements:
if setOfFilesOrDirs:
l.append(heading + ":\n")
l.extend( (" " + str(f) for f in sorted(setOfFilesOrDirs)) )
l.append('') # ensure a blank line follows the list
else:
l.append(heading + ": none")
return '\n'.join(l)
def printReport(self):
title = "{prg} report".format(prg=PROGNAME)
print("\n" + title + '\n' + len(title)*"=", end="\n\n")
print(self.summaryString())
@enum.unique
class FailedCheckReason(enum.Enum):
"""Reasons that can cause 'check' mode to report a mismatch.
Note that network errors and things like that do *not* belong here.
"""
missingDirIndexFile, mismatchingHashForDirIndexFile, \
missingNormalFile, mismatchingHashForNormalFile, \
orphanFile, orphanDirectory = range(6)
# 'path': VirtualPath instance for a file or directory
def explain(self, path):
if self is FailedCheckReason.missingDirIndexFile:
res = ".dirindex file '{}' is missing locally".format(path)
elif self is FailedCheckReason.mismatchingHashForDirIndexFile:
res = ".dirindex file '{}' doesn't have the hash it " \
"should have according to the server".format(path)
elif self is FailedCheckReason.missingNormalFile:
res = "file '{}' is present on the server but missing locally" \
.format(path)
elif self is FailedCheckReason.mismatchingHashForNormalFile:
res = "file '{}' doesn't have the hash given in the " \
".dirindex file of its containing directory".format(path)
elif self is FailedCheckReason.orphanFile:
res = "file '{}' was found locally but is not present on the " \
"server".format(path)
elif self is FailedCheckReason.orphanDirectory:
res = "directory '{}' was found locally but is not present " \
"on the server".format(path)
else:
assert False, "Unhandled enum value: {!r}".format(self)
return res
class TerraSync:
@enum.unique
class Mode(enum.Enum):
"""Main modes of operation for the TerraSync class."""
# Using lower case for the member names, because this way
# enumMember.name is exactly the mode string passed to --mode on the
# command line (can be useful for messages destined to users).
check, sync = range(2)
def __init__(self, mode, doReport, url, target, quick, removeOrphan,
downloadBoundaries):
self.mode = self.Mode[mode]
self.doReport = doReport
self.setUrl(url).setTarget(target)
self.quick = quick
self.removeOrphan = removeOrphan
self.httpGetter = None
self.downloadBoundaries = downloadBoundaries
# Status of the local repository (as compared to what the server says),
# before any update we might do to it.
self.report = Report(self.target)
def inSyncMode(self):
return self.mode == self.Mode.sync
def setUrl(self, url):
self.url = url.rstrip('/').strip()
return self
def setTarget(self, target):
# Using os.path.abspath() here is safer in case the process later uses
# os.chdir(), which would change the meaning of the "." directory.
self.target = os.path.abspath(target)
return self
def start(self, virtualSubdir=VirtualPath('/')):
"""Start the 'sync' or 'check' process.
The 'virtualSubdir' argument must be a VirtualPath instance and
allows one to start the 'sync' or 'check' process in a chosen
subdirectory of the TerraSync repository, instead of at its
root.
"""
# Remove the leading '/' from 'virtualSubdir' and convert to native
# separators ('/' or '\' depending on the platform).
localSubdir = os.path.normpath(virtualSubdir.asRelative())
if localSubdir == ".": # just ugly, but it wouldn't hurt
localSubdir = ""
assert not os.path.isabs(localSubdir), repr(localSubdir)
self.httpGetter = HTTPGetter(self.url)
# Get the hash of the .dirindex file for 'virtualSubdir'
try:
request = HTTPSocketRequest(virtualSubdir / ".dirindex")
with self.httpGetter.get(request) as fileLike:
dirIndexHash = computeHash(fileLike)
except HTTPException as exc:
raise NetworkError("for the root .dirindex file: {errMsg}"
.format(errMsg=exc)) from exc
# Process the chosen part of the repository (recursive)
self.processDirectoryEntry(virtualSubdir, localSubdir, dirIndexHash)
return self.report
def processFileEntry(self, virtualPath, localPath, fileHash):
"""Process a file entry from a .dirindex file."""
localFullPath = join(self.target, localPath)
failedCheckReason = None
if not os.path.isfile(localFullPath):
self.report.addMissingFile(virtualPath)
failedCheckReason = FailedCheckReason.missingNormalFile
elif hashForFile(localFullPath) != fileHash:
self.report.addFileWithMismatchingHash(virtualPath)
failedCheckReason = FailedCheckReason.mismatchingHashForNormalFile
else:
# The file exists and has the hash mentioned in the .dirindex file
return
assert failedCheckReason is not None
if self.inSyncMode():
if os.path.isdir(localFullPath):
# 'localFullPath' is a directory (locally), but on the server
# it is a file -> remove the dir so that we can store the file.
removeDirectoryTree(self.target, localFullPath)
print("Downloading '{}'".format(virtualPath))
request = HTTPDownloadRequest(self, virtualPath, localFullPath)
self.httpGetter.get(request)
else:
self.abortCheckMode(failedCheckReason, virtualPath)
def processDirectoryEntry(self, virtualPath, localPath, dirIndexHash):
"""Process a directory entry from a .dirindex file."""
print("Processing '{}'...".format(virtualPath))
isOuterBucket = True if len(virtualPath.parts) <= 3 else False
coord = parse_terrasync_coordinate(virtualPath.name)
if (coord and
not self.downloadBoundaries.is_coordinate_inside_boundaries(coord, isOuterBucket)):
self.report.addSkippedDueToBoundaries(virtualPath)
return
localFullPath = join(self.target, localPath)
localDirIndex = join(localFullPath, ".dirindex")
failedCheckReason = None
if not os.path.isfile(localDirIndex):
failedCheckReason = FailedCheckReason.missingDirIndexFile
self.report.addMissingDirIndex(virtualPath)
elif hashForFile(localDirIndex) != dirIndexHash:
failedCheckReason = FailedCheckReason.mismatchingHashForDirIndexFile
self.report.addDirIndexWithMismatchingHash(virtualPath)
if failedCheckReason is None:
if not self.quick:
self.handleDirindexFile(localDirIndex)
elif self.inSyncMode():
if not os.path.exists(localFullPath):
os.makedirs(localFullPath)
request = HTTPDownloadRequest(self,
virtualPath / ".dirindex",
localDirIndex,
self.handleDirindexRequest)
self.httpGetter.get(request)
else:
self.abortCheckMode(failedCheckReason, virtualPath / ".dirindex")
def handleDirindexRequest(self, dirindexRequest):
self.handleDirindexFile(dirindexRequest.dst)
def handleDirindexFile(self, dirindexFile):
dirIndex = dirindex.DirIndex(dirindexFile)
virtualBase = dirIndex.path # VirtualPath instance
relativeBase = virtualBase.asRelative() # string, doesn't start with '/'
serverFiles = []
serverDirs = []
for file in dirIndex.files:
f = file['name']
self.processFileEntry(virtualBase / f,
join(relativeBase, f),
file['hash'])
serverFiles.append(f)
for subdir in dirIndex.directories:
d = subdir['name']
self.processDirectoryEntry(virtualBase / d,
join(relativeBase, d),
subdir['hash'])
serverDirs.append(d)
for tarball in dirIndex.tarballs:
# Tarballs are handled the same as normal files.
f = tarball['name']
self.processFileEntry(virtualBase / f,
join(relativeBase, f),
tarball['hash'])
serverFiles.append(f)
localFullPath = join(self.target, relativeBase)
localFiles = [ f for f in listdir(localFullPath)
if isfile(join(localFullPath, f)) ]
for f in localFiles:
if f != ".dirindex" and f not in serverFiles:
virtualPath = virtualBase / f
self.report.addOrphanFile(virtualPath)
if self.inSyncMode():
if self.removeOrphan:
os.remove(join(self.target, virtualPath.asRelative()))
else:
self.abortCheckMode(FailedCheckReason.orphanFile,
virtualPath)
localDirs = [ f for f in listdir(localFullPath)
if isdir(join(localFullPath, f)) ]
for d in localDirs:
if d not in serverDirs:
virtualPath = virtualBase / d
self.report.addOrphanDir(virtualPath)
if self.inSyncMode():
if self.removeOrphan:
removeDirectoryTree(self.target,
join(self.target,
virtualPath.asRelative()))
else:
self.abortCheckMode(FailedCheckReason.orphanDirectory,
virtualPath)
# 'reason' is a member of the FailedCheckReason enum
def abortCheckMode(self, reason, fileOrDirVirtualPath):
assert self.mode == self.Mode.check, repr(self.mode)
print("{prg}: exiting from 'check' mode because {explanation}."
.format(prg=PROGNAME,
explanation=reason.explain(fileOrDirVirtualPath)))
if self.doReport:
self.report.printReport()
sys.exit(ExitStatus.CHECK_MODE_FOUND_MISMATCH.value)
#################################################################################################################################
def parseCommandLine():
parser = argparse.ArgumentParser()
parser.add_argument("-u", "--url", dest="url", metavar="URL",
default="http://flightgear.sourceforge.net/scenery",
help="server URL [default: %(default)s]")
parser.add_argument("-t", "--target", dest="target", metavar="DIR",
default=".", help="""\
directory where to store the files [default: the current directory]""")
parser.add_argument("--only-subdir", dest="onlySubdir", metavar="SUBDIR",
default="", help="""\
restrict processing to this subdirectory of the TerraSync repository. Use
a path relative to the repository root, for instance 'Models/Residential'
[default: process the whole repository]""")
parser.add_argument("-q", "--quick", dest="quick", action="store_true",
default=False, help="enable quick mode")
parser.add_argument("-r", "--remove-orphan", dest="removeOrphan",
action="store_true",
default=False, help="remove old scenery files")
parser.add_argument("--mode", default="sync", choices=("check", "sync"),
help="""\
main mode of operation (default: '%(default)s'). In 'sync' mode, contents
is downloaded from the server to the target directory. On the other hand,
in 'check' mode, {progname} compares the contents of the target directory
with the remote repository without writing nor deleting anything on
disk.""".format(progname=PROGNAME))
parser.add_argument("--report", dest="report", action="store_true",
default=False,
help="""\
before normal exit, print a report of what was found""")
parser.add_argument("--top", dest="top", type=int, default=90, help="""\
maximum latitude to include in download [default: %(default)d]""")
parser.add_argument("--bottom", dest="bottom", type=int, default=-90,
help="""\
minimum latitude to include in download [default: %(default)d]""")
parser.add_argument("--left", dest="left", type=int, default=-180, help="""\
minimum longitude to include in download [default: %(default)d]""")
parser.add_argument("--right", dest="right", type=int, default=179,
help="""\
maximum longitude to include in download [default: %(default)d]""")
args = parser.parse_args()
# Perform consistency checks on the arguments
if args.mode == "check" and args.removeOrphan:
print("{prg}: 'check' mode is read-only and thus doesn't make sense "
"with\noption --remove-orphan (-r)".format(prg=PROGNAME),
file=sys.stderr)
sys.exit(ExitStatus.ERROR.value)
# Replace backslashes with forward slashes, remove leading and trailing
# slashes, collapse consecutive slashes. Yes, this implies that we tolerate
# leading slashes for --only-subdir (which makes sense because virtual
# paths are printed like that by this program, therefore it is natural for
# users to copy & paste such paths in order to use them for --only-subdir).
args.virtualSubdir = VirtualPath(args.onlySubdir.replace('\\', '/'))
# Be nice to our user in case the path starts with '\', 'C:\', etc.
if os.path.isabs(args.virtualSubdir.asRelative()):
print("{prg}: option --only-subdir expects a *relative* path, but got "
"'{subdir}'".format(prg=PROGNAME, subdir=args.onlySubdir),
file=sys.stderr)
sys.exit(ExitStatus.ERROR.value)
return args
def main():
args = parseCommandLine()
terraSync = TerraSync(args.mode, args.report, args.url, args.target,
args.quick, args.removeOrphan,
DownloadBoundaries(args.top, args.left, args.bottom,
args.right))
report = terraSync.start(args.virtualSubdir)
if args.report:
report.printReport()
sys.exit(ExitStatus.SUCCESS.value)