Source code for gavo.helpers.filestuff

File- and directory related helpers for resource utilities.

#c Copyright 2008-2022, the GAVO project <>
#c This program is free software, covered by the GNU GPL.  See the
#c COPYING file in the source distribution.

import os
import re
import warnings

from gavo import base
from gavo import rscdef
from gavo.base import parsecontext

[docs]class Error(base.Error): pass
fnamePat = re.compile(r"([^.]*)(\..*)")
[docs]def stingySplitext(fName): """returns name, extension for fName. The main difference to os.path.splitext is that the main name is not allowed to contain dots and the extension can contain more than one dot. fName is supposed to be a single file name without any path specifier (you might get away with it if your directories do not contain dots, though). """ mat = fnamePat.match(fName) if mat: return, else: return fName, ""
[docs]class FileRenamer(object): """is a name mapper for file rename operations and the like. Warning: This whole thing more or less pretends there are no symlinks. """ def __init__(self, map, showOnly=False):, self.showOnly = map, showOnly
[docs] @classmethod def loadFromFile(cls, inF, **kwargs): """returns a name map for whatever is serialized in the file inF. The format of fName is line-base, with each line being one of - empty -- ignored - beginning with a hash -- ignored - <old> -> <new> -- introducing a map """ map = {} try: for ln in inF: if not ln.strip() or ln.strip().startswith("#"): continue old, new = [s.strip() for s in ln.split("->", 2)] if old in map: raise Error("Two mappings for %s"%old) map[old] = new except ValueError: raise base.ui.logOldExc(Error("Invalid mapping line: %s"%repr(ln))) return cls(map, **kwargs)
[docs] def getFileMap(self, path): """returns a dictionary old->new of renames within path. """ fileMap = {} for dir, subdirs, fNames in os.walk(path): for fName in fNames: stem, ext = stingySplitext(fName) if stem in fileMap[os.path.join(dir, fName)] = os.path.join(dir,[stem]+ext) return fileMap
[docs] def makeRenameProc(self, fileMap): """returns a sequence of (old,new) pairs that, when executed, keep any new from clobbering any existing old. The function will raise an Error if there's a cycle in fileMap. fileMap will be destroyed by this procedure """ proc = [] def addOp(src, dest, sources=None): if sources is None: sources = set() if dest in sources: raise Error("Rename cycle involving %s"%sources) if dest in fileMap: sources.add(src) addOp(dest, fileMap.pop(dest), sources) proc.append((src, dest)) while fileMap: addOp(*fileMap.popitem()) return proc
[docs] def renameInPath(self, path): """performs a name map below path. The rules are: - extensions are ignored -- if we map foo to bar, foo.txt and foo.asc will be renamed bar.txt and foo.txt respectively - the order of statements in the source is irrelevant. However, we try not to clobber anything we've just renamed and will complain about cycles. Also, each file will be renamed not more than once. """ fileMap = self.getFileMap(path) for src, dest in self.makeRenameProc(fileMap): if os.path.exists(dest): raise Error("Request to clobber %s"%repr(dest)) if os.path.exists(src): if self.showOnly: print("%s -> %s"%(src, dest)) else: os.rename(src, dest) else: if not os.path.exists(dest): warnings.warn("Neither source nor dest found in pair %s, %s"%(src, dest))
[docs] def renameInPaths(self, pathList): for path in pathList: self.renameInPath(path)
[docs]def iterSources(ddId, args=[]): """iterates over the current sources of the data descriptor ddId (which is qualified like rdId#id If you pass something nonempty to args, an iterator over its values will be returned. This is for convenient implementation of scripts that work on CL arguments if given, on all files otherwise. """ if args: return iter(args) else: if ddId.count("#")!=1: raise base.ReportableError("iterSources must get a fully qualified id" " (i.e., one with exactly one hash).") dd = parsecontext.resolveCrossId(ddId, rscdef.DataDescriptor) return dd.sources.iterSources()