Source code for armi.utils.textProcessors

# Copyright 2019 TerraPower, LLC
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Utility classes and functions for manipulating text files.
"""
import os
import re
import io
import pathlib
from typing import List, Tuple, Union, Optional, TextIO

from armi import runLog


_INCLUDE_CTOR = False
_INCLUDE_RE = re.compile(r"^([^#]*\s+)?!include\s+(.*)\n?$")
_INDENT_RE = re.compile(r"^[\s\-\?:]*([^\s\-\?:].*)?$")

# String constants
SCIENTIFIC_PATTERN = r"[+-]?\d*\.\d+[eEdD][+-]\d+"
"""
Matches:
* code:` 1.23e10`
* code:`-1.23Ee10`
* code:`+1.23d10`
* code:`  .23D10`
* code:` 1.23e-10`
* code:` 1.23e+1`
"""

FLOATING_PATTERN = r"[+-]?\d+\.*\d*"
"""
Matches 1, 100, 1.0, -1.2, +12.234
"""

DECIMAL_PATTERN = r"[+-]?\d*\.\d+"
"""matches .1, 1.213423, -23.2342, +.023
"""


[docs]class FileMark: def __init__(self, fName, line, column, relativeTo): self.path = fName self.line = line self.column = column # if the path is relative, where is it relative to? We need this to be able to # normalize relative paths to a root file. self.relativeTo = relativeTo def __str__(self): return "{}, line {}, column {}".format(self.path, self.line, self.column)
def _processIncludes( src, out, includes: List[Tuple[pathlib.Path, FileMark]], root, indentation=0, currentFile="<stream>", ): """ This is the workhorse of ``resolveMarkupInclusions`` and friends. Recursively inserts the contents of !included YAML files into the output stream, keeping track of indentation and a list of included files along the way. """ def _beginningOfContent(line: str) -> int: """ Return the position of the first "content" character. This follows the YAML spec at https://yaml.org/spec/current.html#id2519916 In short, it will return the position of the first character that is not whitespace or one of the special "block collection" markers ("-", "?", and ":") """ m = _INDENT_RE.match(line) if m and m.group(1) is not None: return m.start(1) else: return 0 indentSpace = " " * indentation for i, line in enumerate(src.readlines()): leadingSpace = indentSpace if i > 0 else "" m = _INCLUDE_RE.match(line) if m: # this line has an !include on it if m.group(1) is not None: out.write(leadingSpace + m.group(1)) fName = pathlib.Path(os.path.expandvars(m.group(2))) path = root / fName if not path.exists(): raise ValueError( "The !included file, `{}` does not exist from {}!".format( fName, root ) ) includes.append((fName, FileMark(currentFile, i, m.start(2), root))) with open(path, "r") as includedFile: firstCharacterPos = _beginningOfContent(line) newIndent = indentation + firstCharacterPos _processIncludes( includedFile, out, includes, path.parent, indentation=newIndent, currentFile=path, ) else: out.write(leadingSpace + line)
[docs]def resolveMarkupInclusions( src: Union[TextIO, pathlib.Path], root: Optional[pathlib.Path] = None ) -> io.StringIO: r""" Process a text stream, appropriately handling ``!include`` tags. This will take the passed IO stream or file path, replacing any instances of ``!include [path]`` with the appropriate contents of the ``!include`` file. What is returned is a new text stream, containing the contents of all of the files stitched together. Parameters ---------- src : TextIOBase or Path If a Path is provided, read text from there. If is stream is provided, consume text from the stream. If a stream is provided, ``root`` must also be provided. root : Optional Path The root directory to use for resolving relative paths in !include tags. If a stream is provided for ``src``, ``root`` must be provided. Otherwise, the directory containing the ``src`` path will be used by default. Notes ----- While the use of ``!include`` appears as though it would invoke some sort of special custom YAML constructor code, this does not do that. Processing these inclusions as part of the document parsing/composition that comes with ruamel.yaml could work, but has a number of prohibitive drawbacks (or at least reasons why it might not be worth doing). Using a custom constructor is more-or-less supported by ruamel.yaml (which we do use, as it is what underpins the yamlize package), but it carries limitations about how anchors and aliases can cross included-file boundaries. Getting around this requires either monkey-patching ruamel.yaml, or subclassing it, which in turn would require monkey-patching yamlize. Instead, we treat the ``!include``\ s as a sort of pre-processor directive, which essentially pastes the contents of the ``!include``\ d file into the location of the ``!include``. The result is a text stream containing the entire contents, with all ``!include``\ s resolved. The only degree of sophistication lies in how indentation is handled; since YAML cares about indentation to keep track of object hierarchy, care must be taken that the included file contents are indented appropriately. To precisely describe how the indentation works, it helps to have some definitions: - Included file: The file specified in the ``!include [Included file]`` - Including line: The line that actually contains the ``!include [Included file]`` - Meaningful YAML content: Text in a YAML file that is not either indentation or a special character like "-", ":" or "?". The contents of the included file will be indented such that that the first character of each line in the included file will be found at the first column in the including line that contains meaningful YAML content. The only exception is the first line of the included file, which starts at the location of the ``!include`` itself and is not deliberately indented. In the future, we may wish to do the more sophisticated processing of the ``!include``\ s as part of the YAML parse. For future reference, there is some pure gold on that topic here: https://stackoverflow.com/questions/44910886/pyyaml-include-file-and-yaml-aliases-anchors-references """ return _resolveMarkupInclusions(src, root)[0]
def _getRootFromSrc( src: Union[TextIO, pathlib.Path], root: Optional[pathlib.Path] ) -> pathlib.Path: if isinstance(src, pathlib.Path): root = root or src.parent.absolute() elif isinstance(src, io.TextIOBase): if root is None: raise ValueError("A stream was provided without a root directory.") else: raise TypeError("Unsupported source type: `{}`!".format(type(src))) return root
[docs]def findYamlInclusions( src: Union[TextIO, pathlib.Path], root: Optional[pathlib.Path] = None ) -> List[Tuple[pathlib.Path, FileMark]]: """ Return a list containing all of the !included YAML files from a root file. This will attempt to "normalize" relative paths to the passed root. If that is not possible, then an absolute path will be used instead. For example, if a file (A) !includes another file (B) by an absolute path, which in turn !includes more files relative to (B), all of (B)'s relative includes will be turned into absolute paths from the perspective of the root file (A). """ includes = _resolveMarkupInclusions(src, root)[1] root = _getRootFromSrc(src, root) normalizedIncludes = [] for path, mark in includes: if not path.is_absolute(): try: path = (mark.relativeTo / path).relative_to(root or os.getcwd()) except ValueError as _: # Can't make a relative path. IMO, pathlib gives up a little too early, # but we still probably want to decay to absolute paths if the files # arent in the same tree. path = (mark.relativeTo / path).absolute() normalizedIncludes.append((path, mark)) return normalizedIncludes
def _resolveMarkupInclusions( src: Union[TextIO, pathlib.Path], root: Optional[pathlib.Path] = None ) -> Tuple[io.StringIO, List[Tuple[pathlib.Path, FileMark]]]: root = _getRootFromSrc(src, root) if isinstance(src, pathlib.Path): # this is inefficient, but avoids having to play with io buffers with open(src, "r") as rootFile: src = io.StringIO(rootFile.read()) out = io.StringIO() includes = [] _processIncludes(src, out, includes, root) out.seek(0) # be kind; rewind src.seek(0) return out, includes
[docs]class SequentialReader: r""" Fast sequential reader that must be used within a with statement. Attributes ---------- line : str value of the current line match : re.match value of the current match Notes ----- This reader will sequentially search a file for a regular expression pattern or string depending on the method used. When the pattern/string is matched/found, the reader will stop, return :code:`True`, and set the attributes :code:`line` and :code:`match`. This pattern makes it easy to cycle through repetitive output in a very fast manner. For example, if you had a text file with consistent chuncks of information that always started with the same text followed by information, you could do something like this: >>> with SequentialReader('somefile') as sr: ... data = [] ... while sr.searchForText('start of data chunk'): ... # this needs to repeat for as many chunks as there are. ... if sr.searchForPatternOnNextLine('some-(?P<data>\w+)-pattern'): ... data.append(sr.match['data']) """ def __init__(self, filePath): self._filePath = filePath self._stream = None self.line = "" self.match = None self._textErrors = [] self._textWarnings = [] self._patternErrors = [] self.ignoreAllErrors = False
[docs] def issueWarningOnFindingText(self, text, warning): """Add a text search for every line of the file, if the text is found the specified warning will be issued. This is important for determining if issues occurred while searching for text. Parameters ---------- text : str text to find within the file warning : str An warning message to issue. See Also -------- raiseErrorOnFindingText raiseErrorOnFindingPattern """ self._textWarnings.append((text, warning))
[docs] def raiseErrorOnFindingText(self, text, error): """Add a text search for every line of the file, if the text is found the specified error will be raised. This is important for determining if errors occurred while searching for text. Parameters ---------- text : str text to find within the file error : Exception An exception to raise. See Also -------- raiseErrorOnFindingPattern """ self._textErrors.append((text, error))
[docs] def raiseErrorOnFindingPattern(self, pattern, error): """Add a pattern search for every line of the file, if the pattern is found the specified error will be raised. This is important for determining if errors occurred while searching for text. Parameters ---------- pattern : str regular expression pattern error : Exception An exception to raise. See Also -------- raiseErrorOnFindingText """ self._patternErrors.append((re.compile(pattern), error))
def __repr__(self): return "<{} {} {}>".format( self.__class__.__name__, self._filePath, "open" if self._stream is not None else "closed", ) def __enter__(self): if not os.path.exists(self._filePath): raise OSError("Cannot open non-existing file {}".format(self._filePath)) self._stream = open(self._filePath, "r") return self def __exit__(self, exc_type, exc_value, traceback): # if checking for errors, we need to keep reading if ( exc_type is not None and not self.ignoreAllErrors and (self._patternErrors or self._textErrors) ): while self._readLine(): # all lines have '\n' terminators pass if self._stream is not None: try: self._stream.close() except: # pylint: disable=bare-except # We really don't care if anything fails here, plus an exception in exit is ignored anyway pass self._stream = None
[docs] def searchForText(self, text): """Search the file for the next occurrence of :code:`text`, and set the :code:`self.line` attribute to that line's value if it matched. Notes ----- This will search the file line by line until it finds the text. This sets the attribute :code:`self.line`. If the previous :code:`_searchFor*` method did not match, the last line it did not match will be searched first. Returns ------- matched : bool Boolean inidcating whether or not the pattern matched """ self.match = None while True: if text in self.line: return True self.line = self._readLine() if self.line == "": break return False
[docs] def searchForPattern(self, pattern): """Search the file for the next occurece of :code:`pattern` and set the :code:`self.line` attribute to that line's value if it matched. Notes ----- This will search the file line by line until it finds the pattern. This sets the attribute :code:`self.line`. If the previous :code:`_searchFor*` method did not match, the last line it did not match will be searched first. Returns ------- matched : bool Boolean inidcating whether or not the pattern matched """ while True: self.match = re.search(pattern, self.line) if self.match is not None: return True self.line = self._readLine() if self.line == "": break return False
[docs] def searchForPatternOnNextLine(self, pattern): """Search the next line for a given pattern, and set the :code:`self.line` attribute to that line's value if it matched. Notes ----- This sets the attribute :code:`self.line`. If the previous :code:`_searchFor*` method did not match, the last line it did not match will be searched first. Returns ------- matched : bool Boolean inidcating whether or not the pattern matched """ self.match = re.search(pattern, self.line) if self.match is None: self.line = self._readLine() self.match = re.search(pattern, self.line) return self.match is not None
def _readLine(self): line = self._stream.readline() if not self.ignoreAllErrors: for text, error in self._textErrors: if text in line: raise error for text, warning in self._textWarnings: if text in line: runLog.warning(warning) for regex, error in self._patternErrors: if regex.match(line): raise error return line
[docs] def consumeLine(self): """Consumes the line. This is necessary when searching for the same pattern repetitively, because otherwise searchForPatternOnNextLine would not work. """ self.line = "" self.match = None
[docs]class SequentialStringIOReader(SequentialReader): """ Fast sequential reader that must be used within a with statement. Attributes ---------- line : str value of the current line match : re.match value of the current match Notes ----- This reader will sequentially search a file for a regular expression pattern or string depending on the method used. When the pattern/string is matched/found, the reader will stop, return :code:`True`, and set the attributes :code:`line` and :code:`match`. This pattern makes it easy to cycle through repetitive output in a very fast manner. For example, if you had a text file with consistent chuncks of information that always started with the same text followed by information, you could do something like this: >>> with SequentialReader('somefile') as sr: ... data = [] ... while sr.searchForText('start of data chunk'): ... # this needs to repeat for as many chunks as there are. ... if sr.searchForPatternOnNextLine('some-(?P<data>\\w+)-pattern'): ... data.append(sr.match['data']) """ def __init__(self, stringIO): SequentialReader.__init__(self, "StringIO") self._stream = stringIO def __enter__(self): """ Override to prevent trying to open/reopen a StringIO object. We don't need to override :code:`__exit__`, because it doesn't care if closing the object fails. """ return self
[docs]class TextProcessor: """ A general text processing object that extends python's abilities to scan through huge files. Use this instead of a raw file object to read data out of output files, etc. """ scipat = SCIENTIFIC_PATTERN number = FLOATING_PATTERN decimal = DECIMAL_PATTERN def __init__(self, fname, highMem=False): self.eChecking = False # Preserve python 2-like behavior for unit tests that pass None and provide # their own text data (in py2, passing None to abspath yields cwd; py3 raises) self.fpath = os.path.dirname(os.path.abspath(fname or os.getcwd())) f = None if fname is not None: if os.path.exists(fname): f = open(fname) else: # need this not to fail for detecting when RXSUM doesn't exist, etc. # note: Could make it check before instantiating... raise FileNotFoundError(f"{fname} does not exist.") if not highMem: # keep the file on disk, read as necessary self.f = f else: # read all of f into memory and set up a list that remembers where it is. self.f = SmartList(f)
[docs] def reset(self): r"""rewinds the file so you can search through it again""" self.f.seek(0)
def __repr__(self): return "<Text file at {0}>".format(self.f.name)
[docs] def errorChecking(self, checkForErrors): self.eChecking = checkForErrors
[docs] def checkErrors(self, line): pass
[docs] def fsearch(self, pattern, msg=None, killOn=None, textFlag=False): r""" Searches file f for pattern and displays msg when found. Returns line in which pattern is found or FALSE if no pattern is found. Stops searching if finds killOn first If you specify textFlag=True, the search won't use a regular expression (and can't). The basic result is you get less powerful matching capabilities at a huge speedup (10x or so probably, but that's just a guess.) pattern and killOn must be pure text if you do this. """ current = 0 result = "" if textFlag: # fast, text-only mode for line in self.f: if self.eChecking: self.checkErrors(line) if pattern in line: result = line break elif killOn and killOn in line: result = "" break else: result = "" else: # slower regular expression mode cpat = re.compile(pattern) if killOn: kpat = re.compile(killOn) for line in self.f: if self.eChecking: self.checkErrors(line) if killOn: kill = re.search(kpat, line) if kill: # the kill phrase was found first, so die. result = "" break current = re.search(cpat, line) if current: if msg: print(msg) result = line break if not current: result = "" return result
[docs]class SmartList: r"""A list that does stuff like files do i.e. remembers where it was, can seek, etc. Actually this is pretty slow. so much for being smart. nice idea though.""" def __init__(self, f): self.lines = f.readlines() self.position = 0 self.name = f.name self.length = len(self.lines) def __getitem__(self, index): return self.lines[index] def __setitem__(self, index, line): self.lines[index] = line
[docs] def next(self): if self.position >= self.length: self.position = 0 raise StopIteration else: c = self.position self.position += 1 return self.lines[c]
def __iter__(self): return self def __len__(self): return len(self.lines)
[docs] def seek(self, line): self.position = line
[docs] def close(self): pass