运行hdf5文件时Python崩溃

Question

我现在正在切换到使用终端打开python文件。特别是，我正在测试我使用的脚本，以确保所有的事情都相应地运行。

我有一个文件在试图打开时会很麻烦。它包含

from os.path import isfile
import numpy as np
import h5py

def gcPath(basePath,snapNum,chunkNum=0):
    """ Return absolute path to a group catalog HDF5 file (modify as needed). """
    gcPath = basePath + '/groups_%03d/' % snapNum
    filePath1 = gcPath + 'groups_%03d.%d.hdf5' % (snapNum, chunkNum)
    filePath2 = gcPath + 'fof_subhalo_tab_%03d.%d.hdf5' % (snapNum, chunkNum)

    if isfile(filePath1):
        return filePath1
    return filePath2

def offsetPath(basePath, snapNum):
    """ Return absolute path to a separate offset file (modify as needed). """
    offsetPath = basePath + '../postprocessing/offsets/offsets_%03d.hdf5' % snapNum

    return offsetPath

def loadObjects(basePath,snapNum,gName,nName,fields):
    """ Load either halo or subhalo information from the group catalog. """
    result = {}

    # make sure fields is not a single element
    if isinstance(fields, basestring):
        fields = [fields]

    # load header from first chunk
    with h5py.File(gcPath(basePath,snapNum),'r') as f:

        header = dict( f['Header'].attrs.items() )
        result['count'] = f['Header'].attrs['N'+nName+'_Total']

        if not result['count']:
            print 'warning: zero groups, empty return (snap='+str(snapNum)+').'
            return result

        # if fields not specified, load everything
        if not fields:
            fields = f[gName].keys()

        for field in fields:
            # verify existence
            if not field in f[gName].keys():
                raise Exception("Group catalog does not have requested field ["+field+"]!")

            # replace local length with global
            shape = list(f[gName][field].shape)
            shape[0] = result['count']

            # allocate within return dict
            result[field] = np.zeros( shape, dtype=f[gName][field].dtype )

    # loop over chunks
    wOffset = 0

    for i in range(header['NumFiles']):
        f = h5py.File(gcPath(basePath,snapNum,i),'r')

        if not f['Header'].attrs['N'+nName+'_ThisFile']:
            continue # empty file chunk

        # loop over each requested field
        for field in fields:
            # shape and type
            shape = f[gName][field].shape

            # read data local to the current file
            if len(shape) == 1:
                result[field][wOffset:wOffset+shape[0]] = f[gName][field][0:shape[0]]
            else:
                result[field][wOffset:wOffset+shape[0],:] = f[gName][field][0:shape[0],:]


        wOffset += shape[0]
        f.close()

    # only a single field? then return the array instead of a single item dict
    if len(fields) == 1:
        return result[fields[0]]

    return result

def loadSubhalos(basePath,snapNum,fields=None):
    """ Load all subhalo information from the entire group catalog for one snapshot
       (optionally restrict to a subset given by fields). """

    return loadObjects(basePath,snapNum,"Subhalo","subgroups",fields)

def loadHalos(basePath,snapNum,fields=None):
    """ Load all halo information from the entire group catalog for one snapshot
       (optionally restrict to a subset given by fields). """

    return loadObjects(basePath,snapNum,"Group","groups",fields)

def loadHeader(basePath,snapNum):
    """ Load the group catalog header. """
    with h5py.File(gcPath(basePath,snapNum),'r') as f:
        header = dict( f['Header'].attrs.items() )

    return header

def load(basePath,snapNum):
    """ Load complete group catalog all at once. """
    r = {}
    r['subhalos'] = loadSubhalos(basePath,snapNum)
    r['halos']    = loadHalos(basePath,snapNum)
    r['header']   = loadHeader(basePath,snapNum)
    return r

def loadSingle(basePath,snapNum,haloID=-1,subhaloID=-1):
    """ Return complete group catalog information for one halo or subhalo. """
    if (haloID < 0 and subhaloID < 0) or (haloID >= 0 and subhaloID >= 0):
        raise Exception("Must specify either haloID or subhaloID (and not both).")

    gName = "Subhalo" if subhaloID >= 0 else "Group"
    searchID = subhaloID if subhaloID >= 0 else haloID

     # old or new format
    if 'fof_subhalo' in gcPath(basePath,snapNum):
        # use separate 'offsets_nnn.hdf5' files
        with h5py.File(offsetPath(basePath,snapNum),'r') as f:
             offsets = f['FileOffsets/'+gName][()]
    else:
        # use header of group catalog
        with h5py.File(gcPath(basePath,snapNum),'r') as f:
        offsets = f['Header'].attrs['FileOffsets_'+gName]

    offsets = searchID - offsets
    fileNum = np.max( np.where(offsets >= 0) )
    groupOffset = offsets[fileNum]

    # load halo/subhalo fields into a dict
    result = {}

    with h5py.File(gcPath(basePath,snapNum,fileNum),'r') as f:
        for haloProp in f[gName].keys():
            result[haloProp] = f[gName][haloProp][groupOffset]

    return result

基本上，这个文件所做的是从我的用户目录中检索定义的hdf5文件进行分析。

如果我要在python环境中运行它，那么返回时

^{pr2}$

我以前使用过jupyter，它有一个预先安装的包的大列表，所以问题可能是我没有安装特定的包。但是看看回溯错误，问题似乎是^{cd1>}。但是已经安装了^{cd1>}。

Answer 1

您的示例代码相当长。一个更简洁的例子会使帮助更容易。在

对于HDF5错误，通常最重要的错误首先出现，这里是“无法初始化转换函数”。HDF5文件中有什么类型的数据？数据类型是否可以映射到？在