-# Content DB
+# ContentDB
# Copyright (C) 2018 rubenwardy
#
# This program is free software: you can redistribute it and/or modify
# along with this program. If not, see <https://www.gnu.org/licenses/>.
-import flask, json, os, git, tempfile, shutil
+import flask, json, os, git, tempfile, shutil, gitdb
from git import GitCommandError
-from flask.ext.sqlalchemy import SQLAlchemy
+from git_archive_all import GitArchiver
+from flask_sqlalchemy import SQLAlchemy
from urllib.error import HTTPError
import urllib.request
from urllib.parse import urlparse, quote_plus, urlsplit
+from zipfile import ZipFile
+
from app import app
from app.models import *
from app.tasks import celery, TaskError
-from app.utils import randomString
-
-
-class GithubURLMaker:
- def __init__(self, url):
- # Rewrite path
- import re
- m = re.search("^\/([^\/]+)\/([^\/]+)\/?$", url.path)
- if m is None:
- return
-
- user = m.group(1)
- repo = m.group(2).replace(".git", "")
- self.baseUrl = "https://raw.githubusercontent.com/{}/{}/master" \
- .format(user, repo)
- self.user = user
- self.repo = repo
-
- def isValid(self):
- return self.baseUrl is not None
-
- def getRepoURL(self):
- return "https://github.com/{}/{}".format(self.user, self.repo)
-
- def getScreenshotURL(self):
- return self.baseUrl + "/screenshot.png"
+from app.utils import randomString, getExtension
+from .minetestcheck import build_tree, MinetestCheckError, ContentType
+from .minetestcheck.config import parse_conf
- def getCommitsURL(self, branch):
- return "https://api.github.com/repos/{}/{}/commits?sha={}" \
- .format(self.user, self.repo, urllib.parse.quote_plus(branch))
-
- def getCommitDownload(self, commit):
- return "https://github.com/{}/{}/archive/{}.zip" \
- .format(self.user, self.repo, commit)
krock_list_cache = None
krock_list_cache_by_name = None
global krock_list_cache_by_name
if krock_list_cache is None:
- contents = urllib.request.urlopen("http://krock-works.16mb.com/MTstuff/modList.php").read().decode("utf-8")
+ contents = urllib.request.urlopen("https://krock-works.uk.to/minetest/modList.php").read().decode("utf-8")
list = json.loads(contents)
def h(x):
return krock_list_cache, krock_list_cache_by_name
+
def findModInfo(author, name, link):
list, lookup = getKrockList()
return None
-def parseConf(string):
- retval = {}
- for line in string.split("\n"):
- idx = line.find("=")
- if idx > 0:
- key = line[:idx].strip()
- value = line[idx+1:].strip()
- retval[key] = value
-
- return retval
-
-
-class PackageTreeNode:
- def __init__(self, baseDir, author=None, repo=None, name=None):
- print("Scanning " + baseDir)
- self.baseDir = baseDir
- self.author = author
- self.name = name
- self.repo = repo
- self.meta = None
- self.children = []
-
- # Detect type
- type = None
- is_modpack = False
- if os.path.isfile(baseDir + "/game.conf"):
- type = PackageType.GAME
- elif os.path.isfile(baseDir + "/init.lua"):
- type = PackageType.MOD
- elif os.path.isfile(baseDir + "/modpack.txt"):
- type = PackageType.MOD
- is_modpack = True
- elif os.path.isdir(baseDir + "/mods"):
- type = PackageType.GAME
- elif os.listdir(baseDir) == []:
- # probably a submodule
- return
- else:
- raise TaskError("Unable to detect package type!")
-
- self.type = type
- self.readMetaFiles()
-
- if self.type == PackageType.GAME:
- self.addChildrenFromModDir(baseDir + "/mods")
- elif is_modpack:
- self.addChildrenFromModDir(baseDir)
-
-
- def readMetaFiles(self):
- result = {}
-
- # .conf file
- try:
- with open(self.baseDir + "/mod.conf", "r") as myfile:
- conf = parseConf(myfile.read())
- for key in ["name", "description", "title", "depends", "optional_depends"]:
- try:
- result[key] = conf[key]
- except KeyError:
- pass
- except IOError:
- print("description.txt does not exist!")
-
- # description.txt
- if not "description" in result:
- try:
- with open(self.baseDir + "/description.txt", "r") as myfile:
- result["description"] = myfile.read()
- except IOError:
- print("description.txt does not exist!")
-
- # depends.txt
- import re
- pattern = re.compile("^([a-z0-9_]+)\??$")
- if not "depends" in result and not "optional_depends" in result:
- try:
- with open(self.baseDir + "/depends.txt", "r") as myfile:
- contents = myfile.read()
- soft = []
- hard = []
- for line in contents.split("\n"):
- line = line.strip()
- if pattern.match(line):
- if line[len(line) - 1] == "?":
- soft.append( line[:-1])
- else:
- hard.append(line)
-
- result["depends"] = hard
- result["optional_depends"] = soft
-
- except IOError:
- print("depends.txt does not exist!")
-
- else:
- if "depends" in result:
- result["depends"] = [x.strip() for x in result["depends"].split(",")]
- if "optional_depends" in result:
- result["optional_depends"] = [x.strip() for x in result["optional_depends"].split(",")]
-
-
- # Calculate Title
- if "name" in result and not "title" in result:
- result["title"] = result["name"].replace("_", " ").title()
-
- # Calculate short description
- if "description" in result:
- desc = result["description"]
- idx = desc.find(".") + 1
- cutIdx = min(len(desc), 200 if idx < 5 else idx)
- result["short_description"] = desc[:cutIdx]
-
- # Get forum ID
- info = findModInfo(self.author, result.get("name"), self.repo)
- if info is not None:
- result["forumId"] = info.get("topicId")
-
- if "name" in result:
- self.name = result["name"]
- del result["name"]
-
- self.meta = result
-
- def addChildrenFromModDir(self, dir):
- for entry in next(os.walk(dir))[1]:
- path = dir + "/" + entry
- if not entry.startswith('.') and os.path.isdir(path):
- self.children.append(PackageTreeNode(path, name=entry))
-
-
- def fold(self, attr, key=None, acc=None):
- if acc is None:
- acc = set()
-
- if self.meta is None:
- return acc
-
- at = getattr(self, attr)
- value = at if key is None else at.get(key)
-
- if isinstance(value, list):
- acc |= set(value)
- elif value is not None:
- acc.add(value)
-
- for child in self.children:
- child.fold(attr, key, acc)
-
- return acc
-
- def get(self, key):
- return self.meta.get(key)
-
def generateGitURL(urlstr):
scheme, netloc, path, query, frag = urlsplit(urlstr)
return "http://:@" + netloc + path + query
+
+def getTempDir():
+ return os.path.join(tempfile.gettempdir(), randomString(10))
+
+
# Clones a repo from an unvalidated URL.
# Returns a tuple of path and repo on sucess.
# Throws `TaskError` on failure.
# Caller is responsible for deleting returned directory.
def cloneRepo(urlstr, ref=None, recursive=False):
- gitDir = tempfile.gettempdir() + "/" + randomString(10)
+ gitDir = getTempDir()
err = None
try:
gitUrl = generateGitURL(urlstr)
print("Cloning from " + gitUrl)
- repo = git.Repo.clone_from(gitUrl, gitDir, \
- progress=None, env=None, depth=1, recursive=recursive, kill_after_timeout=15)
- if ref is not None:
- repo.create_head("myhead", ref).checkout()
+ if ref is None:
+ repo = git.Repo.clone_from(gitUrl, gitDir, \
+ progress=None, env=None, depth=1, recursive=recursive, kill_after_timeout=15)
+ else:
+ repo = git.Repo.init(gitDir)
+ origin = repo.create_remote("origin", url=gitUrl)
+ assert origin.exists()
+ origin.fetch()
+ origin.pull(ref)
+
+ for submodule in repo.submodules:
+ submodule.update(init=True)
+
return gitDir, repo
+
except GitCommandError as e:
# This is needed to stop the backtrace being weird
err = e.stderr
+ except gitdb.exc.BadName as e:
+ err = "Unable to find the reference " + (ref or "?") + "\n" + e.stderr
+
raise TaskError(err.replace("stderr: ", "") \
.replace("Cloning into '" + gitDir + "'...", "") \
.strip())
+
@celery.task()
def getMeta(urlstr, author):
gitDir, _ = cloneRepo(urlstr, recursive=True)
- tree = PackageTreeNode(gitDir, author=author, repo=urlstr)
+
+ try:
+ tree = build_tree(gitDir, author=author, repo=urlstr)
+ except MinetestCheckError as err:
+ raise TaskError(str(err))
+
shutil.rmtree(gitDir)
result = {}
result["name"] = tree.name
- result["provides"] = tree.fold("name")
+ result["provides"] = tree.getModNames()
result["type"] = tree.type.name
for key in ["depends", "optional_depends"]:
return result
-def makeVCSReleaseFromGithub(id, branch, release, url):
- urlmaker = GithubURLMaker(url)
- if not urlmaker.isValid():
- raise TaskError("Invalid github repo URL")
+def postReleaseCheckUpdate(self, release, path):
+ try:
+ tree = build_tree(path, expected_type=ContentType[release.package.type.name], \
+ author=release.package.author.username, name=release.package.name)
- commitsURL = urlmaker.getCommitsURL(branch)
- contents = urllib.request.urlopen(commitsURL).read().decode("utf-8")
- commits = json.loads(contents)
+ cache = {}
+ def getMetaPackages(names):
+ return [ MetaPackage.GetOrCreate(x, cache) for x in names ]
- if len(commits) == 0 or not "sha" in commits[0]:
- raise TaskError("No commits found")
+ provides = tree.getModNames()
- release.url = urlmaker.getCommitDownload(commits[0]["sha"])
- print(release.url)
- release.task_id = None
- db.session.commit()
+ package = release.package
+ package.provides.clear()
+ package.provides.extend(getMetaPackages(tree.getModNames()))
- return release.url
+ # Delete all meta package dependencies
+ package.dependencies.filter(Dependency.meta_package != None).delete()
+ # Get raw dependencies
+ depends = tree.fold("meta", "depends")
+ optional_depends = tree.fold("meta", "optional_depends")
+ # Filter out provides
+ for mod in provides:
+ depends.discard(mod)
+ optional_depends.discard(mod)
-@celery.task()
-def makeVCSRelease(id, branch):
+ # Add dependencies
+
+ for meta in getMetaPackages(depends):
+ db.session.add(Dependency(package, meta=meta, optional=False))
+
+ for meta in getMetaPackages(optional_depends):
+ db.session.add(Dependency(package, meta=meta, optional=True))
+
+ # Update min/max
+
+ if tree.meta.get("min_minetest_version"):
+ release.min_rel = MinetestRelease.get(tree.meta["min_minetest_version"], None)
+
+ if tree.meta.get("max_minetest_version"):
+ release.max_rel = MinetestRelease.get(tree.meta["max_minetest_version"], None)
+
+ return tree
+
+ except MinetestCheckError as err:
+ if "Fails validation" not in release.title:
+ release.title += " (Fails validation)"
+
+ db.session.rollback()
+ release.task_id = self.request.id
+ release.approved = False
+ db.session.commit()
+
+ raise TaskError(str(err))
+
+
+@celery.task(bind=True)
+def updateMetaFromRelease(self, id, path):
release = PackageRelease.query.get(id)
if release is None:
raise TaskError("No such release!")
elif release.package is None:
raise TaskError("No package attached to release")
- urlmaker = None
- url = urlparse(release.package.repo)
- if url.netloc == "github.com":
- return makeVCSReleaseFromGithub(id, branch, release, url)
- else:
- gitDir, repo = cloneRepo(release.package.repo, ref=branch, recursive=True)
+ print("updateMetaFromRelease: {} for {}/{}" \
+ .format(id, release.package.author.display_name, release.package.name))
- try:
- filename = randomString(10) + ".zip"
- destPath = os.path.join("app/public/uploads", filename)
- with open(destPath, "wb") as fp:
- repo.archive(fp)
+ temp = getTempDir()
+ try:
+ with ZipFile(path, 'r') as zip_ref:
+ zip_ref.extractall(temp)
+
+ postReleaseCheckUpdate(self, release, temp)
+ db.session.commit()
+
+ finally:
+ shutil.rmtree(temp)
+
+
+@celery.task(bind=True)
+def checkZipRelease(self, id, path):
+ release = PackageRelease.query.get(id)
+ if release is None:
+ raise TaskError("No such release!")
+ elif release.package is None:
+ raise TaskError("No package attached to release")
+
+ temp = getTempDir()
+ try:
+ with ZipFile(path, 'r') as zip_ref:
+ zip_ref.extractall(temp)
+
+ postReleaseCheckUpdate(self, release, temp)
- release.url = "/uploads/" + filename
- print(release.url)
- release.task_id = None
- db.session.commit()
+ release.task_id = None
+ release.approve(release.package.author)
+ db.session.commit()
+
+ finally:
+ shutil.rmtree(temp)
+
+
+@celery.task(bind=True)
+def makeVCSRelease(self, id, branch):
+ release = PackageRelease.query.get(id)
+ if release is None:
+ raise TaskError("No such release!")
+ elif release.package is None:
+ raise TaskError("No package attached to release")
+
+ gitDir, repo = cloneRepo(release.package.repo, ref=branch, recursive=True)
+
+ postReleaseCheckUpdate(self, release, gitDir)
+
+ try:
+ filename = randomString(10) + ".zip"
+ destPath = os.path.join(app.config["UPLOAD_DIR"], filename)
+
+ assert(not os.path.isfile(destPath))
+ archiver = GitArchiver(force_sub=True, main_repo_abspath=gitDir)
+ archiver.create(destPath)
+ assert(os.path.isfile(destPath))
+
+ release.url = "/uploads/" + filename
+ release.task_id = None
+ release.commit_hash = repo.head.object.hexsha
+ release.approve(release.package.author)
+ db.session.commit()
+
+ updateMetaFromRelease.delay(release.id, destPath)
+
+ return release.url
+ finally:
+ shutil.rmtree(gitDir)
- return release.url
- finally:
- shutil.rmtree(gitDir)
@celery.task()
def importRepoScreenshot(id):
sourcePath = gitDir + "/screenshot." + ext
if os.path.isfile(sourcePath):
filename = randomString(10) + "." + ext
- destPath = os.path.join("app/public/uploads", filename)
+ destPath = os.path.join(app.config["UPLOAD_DIR"], filename)
shutil.copyfile(sourcePath, destPath)
ss = PackageScreenshot()
return None
-
-def getDepends(package):
- url = urlparse(package.repo)
- urlmaker = None
- if url.netloc == "github.com":
- urlmaker = GithubURLMaker(url)
- else:
- return {}
-
- result = {}
- if not urlmaker.isValid():
- return {}
-
- #
- # Try getting depends on mod.conf
- #
- try:
- contents = urllib.request.urlopen(urlmaker.getModConfURL()).read().decode("utf-8")
- conf = parseConf(contents)
- for key in ["depends", "optional_depends"]:
- try:
- result[key] = conf[key]
- except KeyError:
- pass
-
- except HTTPError:
- print("mod.conf does not exist")
-
- if "depends" in result or "optional_depends" in result:
- return result
-
-
- #
- # Try depends.txt
- #
- import re
- pattern = re.compile("^([a-z0-9_]+)\??$")
- try:
- contents = urllib.request.urlopen(urlmaker.getDependsURL()).read().decode("utf-8")
- soft = []
- hard = []
- for line in contents.split("\n"):
- line = line.strip()
- if pattern.match(line):
- if line[len(line) - 1] == "?":
- soft.append( line[:-1])
- else:
- hard.append(line)
-
- result["depends"] = ",".join(hard)
- result["optional_depends"] = ",".join(soft)
- except HTTPError:
- print("depends.txt does not exist")
-
- return result
-
-
-def importDependencies(package, mpackage_cache):
- if Dependency.query.filter_by(depender=package).count() != 0:
+@celery.task(bind=True)
+def importForeignDownloads(self, id):
+ release = PackageRelease.query.get(id)
+ if release is None:
+ raise TaskError("No such release!")
+ elif release.package is None:
+ raise TaskError("No package attached to release")
+ elif not release.url.startswith("http"):
return
- result = getDepends(package)
-
- if "depends" in result:
- deps = Dependency.SpecToList(package, result["depends"], mpackage_cache)
- print("{} hard: {}".format(len(deps), result["depends"]))
- for dep in deps:
- dep.optional = False
- db.session.add(dep)
-
- if "optional_depends" in result:
- deps = Dependency.SpecToList(package, result["optional_depends"], mpackage_cache)
- print("{} soft: {}".format(len(deps), result["optional_depends"]))
- for dep in deps:
- dep.optional = True
- db.session.add(dep)
-
-@celery.task()
-def importAllDependencies():
- Dependency.query.delete()
- mpackage_cache = {}
- packages = Package.query.filter_by(type=PackageType.MOD).all()
- for i, p in enumerate(packages):
- print("============= {} ({}/{}) =============".format(p.name, i, len(packages)))
- importDependencies(p, mpackage_cache)
-
- db.session.commit()
+ try:
+ ext = getExtension(release.url)
+ filename = randomString(10) + "." + ext
+ filepath = os.path.join(app.config["UPLOAD_DIR"], filename)
+ urllib.request.urlretrieve(release.url, filepath)
+
+ release.url = "/uploads/" + filename
+ db.session.commit()
+
+ except urllib.error.URLError:
+ db.session.rollback()
+ release.task_id = self.request.id
+ release.approved = False
+ db.session.commit()
projects / cheatdb.git / blobdiff