Source code for camcops_server.cc_modules.cc_spreadsheet

"""
camcops_server/cc_modules/cc_spreadsheet.py

===============================================================================

    Copyright (C) 2012, University of Cambridge, Department of Psychiatry.
    Created by Rudolf Cardinal (rnc1001@cam.ac.uk).

    This file is part of CamCOPS.

    CamCOPS 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 3 of the License, or
    (at your option) any later version.

    CamCOPS 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 CamCOPS. If not, see <https://www.gnu.org/licenses/>.

===============================================================================

**Helper functions/classes for spreadsheet-style tab-separated value (TSV)
(and related) exports.**

"""

from collections import OrderedDict
import csv
import io
import logging
import os
import random
import re
from typing import (
    Any,
    BinaryIO,
    Callable,
    Container,
    Dict,
    Iterable,
    List,
    Optional,
    Sequence,
    Union,
)
import zipfile

from cardinal_pythonlib.datetimefunc import (
    format_datetime,
    get_now_localtz_pendulum,
)
from cardinal_pythonlib.excel import (
    convert_for_openpyxl,
    convert_for_pyexcel_ods3,
)
from cardinal_pythonlib.logs import BraceStyleAdapter
from sqlalchemy.engine import CursorResult

from camcops_server.cc_modules.cc_constants import DateFormat

ODS_VIA_PYEXCEL = True  # significantly faster
XLSX_VIA_PYEXCEL = True

if ODS_VIA_PYEXCEL:
    import pyexcel_ods3  # e.g. pip install pyexcel-ods3==0.5.3

    ODSWriter = ODSSheet = None
else:
    from odswriter import ODSWriter, Sheet as ODSSheet

    pyexcel_ods3 = None

if XLSX_VIA_PYEXCEL:
    import pyexcel_xlsx  # e.g. pip install pyexcel-xlsx==0.5.7

    openpyxl = XLWorkbook = XLWorksheet = None
else:
    from openpyxl.workbook.workbook import Workbook as XLWorkbook
    from openpyxl.worksheet.worksheet import Worksheet as XLWorksheet

    pyexcel_xlsx = None

log = BraceStyleAdapter(logging.getLogger(__name__))


# =============================================================================
# Spreadsheet output holding structures
# =============================================================================


[docs]class SpreadsheetPage(object): """ Represents a single "spreadsheet" page, e.g. for TSV/Excel/ODS output. """
[docs] def __init__( self, name: str, rows: List[Union[Dict[str, Any], OrderedDict]] ) -> None: """ Args: name: name for the whole sheet rows: list of rows, where each row is a dictionary mapping column name to value """ assert name, "Missing name" self.name = name self.rows = rows self.headings = [] # type: List[str] for row in rows: self._add_headings_if_absent(row.keys())
def __str__(self) -> str: return f"SpreadsheetPage: name={self.name}\n{self.get_tsv()}"
[docs] @classmethod def from_headings_rows( cls, name: str, headings: List[str], rows: List[Sequence[Any]] ) -> "SpreadsheetPage": """ Creates a SpreadsheetPage object using a list of headings and the row data as a list of lists. """ page = cls(name=name, rows=[]) n_cols = len(headings) page.headings = headings for row in rows: assert len(row) == n_cols page.rows.append(dict(zip(headings, row))) return page
[docs] @classmethod def from_result(cls, name: str, rp: CursorResult) -> "SpreadsheetPage": """ Creates a SpreadsheetPage object from an SQLAlchemy Result. Args: rp: A :class:` sqlalchemy.engine.Result`. name: Name for this sheet. """ column_names = rp.keys() rows = rp.fetchall() return cls.from_headings_rows( name=name, headings=column_names, rows=rows )
@property def empty(self) -> bool: """ Do we have zero rows? """ return len(self.rows) == 0 def _add_headings_if_absent(self, headings: Iterable[str]) -> None: """ Add any headings we've not yet seen to our list of headings. """ for h in headings: if h not in self.headings: self.headings.append(h)
[docs] def add_or_set_value(self, heading: str, value: Any) -> None: """ If we contain only a single row, this function will set the value for a given column (``heading``) to ``value``. Raises: :exc:`AssertionError` if we don't have exactly 1 row """ assert len(self.rows) == 1, "add_value can only be used if #rows == 1" self._add_headings_if_absent([heading]) self.rows[0][heading] = value
[docs] def add_or_set_column(self, heading: str, values: List[Any]) -> None: """ Set the column labelled ``heading`` so it contains the values specified in ``values``. The length of ``values`` must equal the number of rows that we already contain. Raises: :exc:`AssertionError` if the number of values doesn't match the number of existing rows """ assert len(values) == len(self.rows), "#values != #existing rows" self._add_headings_if_absent([heading]) for i, row in enumerate(self.rows): row[heading] = values[i]
[docs] def add_or_set_columns_from_page(self, other: "SpreadsheetPage") -> None: """ This function presupposes that ``self`` and ``other`` are two pages ("spreadsheets") with *matching* rows. It updates values or creates columns in ``self`` such that the values from all columns in ``other`` are written to the corresponding rows of ``self``. Raises: :exc:`AssertionError` if the two pages (sheets) don't have the same number of rows. """ assert len(self.rows) == len(other.rows), "Mismatched #rows" self._add_headings_if_absent(other.headings) for i, row in enumerate(self.rows): for k, v in other.rows[i].items(): row[k] = v
[docs] def add_rows_from_page(self, other: "SpreadsheetPage") -> None: """ Add all rows from ``other`` to ``self``. """ self._add_headings_if_absent(other.headings) self.rows.extend(other.rows)
[docs] def sort_headings(self) -> None: """ Sort our headings internally. """ self.headings.sort()
[docs] def delete_columns(self, headings: Container[str]) -> None: """ Removes columns with the specified heading names. Used to simplify spreadsheets. Since our rows are a dictionary, and our export functions are based on the headings, all we have to do is to delete the unwanted headings. """ self.headings = [h for h in self.headings if h not in headings]
@property def plainrows(self) -> List[List[Any]]: """ Returns a list of rows, where each row is a list of values. Does not include a "header" row. Compare :attr:`rows`, which is a list of dictionaries. """ rows = [] for row in self.rows: rows.append([row.get(h) for h in self.headings]) return rows
[docs] def spreadsheetrows( self, converter: Callable[[Any], Any] ) -> List[List[Any]]: """ Like :meth:`plainrows`, but (a) ensures every cell is converted to a value that can be sent to a spreadsheet converted (e.g. ODS, XLSX), and (b) includes a header row. """ rows = [self.headings.copy()] for row in self.rows: rows.append([converter(row.get(h)) for h in self.headings]) return rows
[docs] def get_tsv(self, dialect: str = "excel-tab") -> str: r""" Returns the entire page (sheet) as TSV: one header row and then lots of data rows. For the dialect, see https://docs.python.org/3/library/csv.html#csv.excel_tab. For CSV files, see RGC 4180: https://tools.ietf.org/html/rfc4180. For TSV files, see https://www.iana.org/assignments/media-types/text/tab-separated-values. Test code: .. code-block:: python import io import csv from typing import List def test(row: List[str], dialect: str = "excel-tab") -> str: f = io.StringIO() writer = csv.writer(f, dialect=dialect) writer.writerow(row) return f.getvalue() test(["hello", "world"]) test(["hello\ttab", "world"]) # actual tab within double quotes test(["hello\nnewline", "world"]) # actual newline within double quotes test(['hello"doublequote', "world"]) # doubled double quote within double quotes """ # noqa f = io.StringIO() writer = csv.writer(f, dialect=dialect) writer.writerow(self.headings) for row in self.rows: writer.writerow([row.get(h) for h in self.headings]) return f.getvalue()
[docs] def write_to_openpyxl_xlsx_worksheet(self, ws: "XLWorksheet") -> None: """ Writes data from this page to an existing ``openpyxl`` XLSX worksheet. """ ws.append(self.headings) for row in self.rows: ws.append( [convert_for_openpyxl(row.get(h)) for h in self.headings] )
[docs] def write_to_odswriter_ods_worksheet(self, ws: "ODSSheet") -> None: """ Writes data from this page to an existing ``odswriter`` ODS sheet. """ # noinspection PyUnresolvedReferences ws.writerow(self.headings) for row in self.rows: # noinspection PyUnresolvedReferences ws.writerow([row.get(h) for h in self.headings])
[docs] def r_object_name(self) -> str: """ Name of the object when imported into R. The main thing: no leading underscores. """ n = self.name n = n[1:] if n.startswith("_") else n return f"camcops_{n}" # less chance of conflict within R
[docs] def r_data_table_definition(self) -> str: """ Returns a string to define this object as a ``data.table`` in R. See also: - https://stackoverflow.com/questions/32103639/read-csv-file-in-r-with-double-quotes """ # noqa object_name = self.r_object_name() csv_text = self.get_tsv(dialect="excel") csv_text = csv_text.replace('"', r"\"") definition = ( f'data.table::fread(sep=",", header=TRUE, text="{csv_text}"\n)' ) return f"{object_name} <- {definition}"
[docs]class SpreadsheetCollection(object): """ A collection of :class:`camcops_server.cc_modules.cc_spreadsheet.SpreadsheetPage` pages (spreadsheets), like an Excel workbook. """
[docs] def __init__(self) -> None: self.pages = [] # type: List[SpreadsheetPage]
def __str__(self) -> str: return "SpreadsheetCollection:\n" + "\n\n".join( page.get_tsv() for page in self.pages ) # ------------------------------------------------------------------------- # Pages # -------------------------------------------------------------------------
[docs] def page_with_name(self, page_name: str) -> Optional[SpreadsheetPage]: """ Returns the page with the specific name, or ``None`` if no such page exists. """ return next( (page for page in self.pages if page.name == page_name), None )
[docs] def add_page(self, page: SpreadsheetPage) -> None: """ Adds a new page to our collection. If the new page has the same name as an existing page, rows from the new page are added to the existing page. Does nothing if the new page is empty. """ if page.empty: return existing_page = self.page_with_name(page.name) if existing_page: # Blend with existing page existing_page.add_rows_from_page(page) else: # New page self.pages.append(page)
[docs] def add_pages(self, pages: List[SpreadsheetPage]) -> None: """ Adds all ``pages`` to our collection, via :func:`add_page`. """ for page in pages: self.add_page(page)
[docs] def sort_headings_within_all_pages(self) -> None: """ Sort headings within each of our pages. """ for page in self.pages: page.sort_headings()
[docs] def sort_pages(self) -> None: """ Sort our pages by their page name. """ self.pages.sort(key=lambda p: p.name)
[docs] def get_page_names(self) -> List[str]: """ Return a list of the names of all our pages. """ return [p.name for p in self.pages]
[docs] def delete_page(self, page_name: str) -> None: """ Delete any page with the name specified. """ self.pages = [p for p in self.pages if p.name != page_name]
[docs] def delete_pages(self, page_names: Container[str]) -> None: """ Delete pages with the names specified. """ self.pages = [p for p in self.pages if p.name not in page_names]
[docs] def delete_columns(self, headings: Container[str]) -> None: """ Across all pages, removes columns with the specified heading names. Used to simplify spreadsheets. """ for p in self.pages: p.delete_columns(headings)
# ------------------------------------------------------------------------- # TSV # -------------------------------------------------------------------------
[docs] def get_tsv_file(self, page_name: str) -> str: """ Returns a TSV file for a named page. Raises: :exc:`AssertionError` if the named page does not exist """ page = self.page_with_name(page_name) assert page is not None, f"No such page with name {page_name}" return page.get_tsv()
# ------------------------------------------------------------------------- # ZIP of TSVs # -------------------------------------------------------------------------
[docs] def write_zip( self, file: Union[str, BinaryIO], encoding: str = "utf-8", compression: int = zipfile.ZIP_DEFLATED, ) -> None: """ Writes data to a file, as a ZIP file of TSV files. Args: file: filename or file-like object encoding: encoding to use when writing the TSV files compression: compression method to use Choice of compression method: see - https://docs.python.org/3/library/zipfile.html - https://pkware.cachefly.net/webdocs/casestudies/APPNOTE.TXT - https://en.wikipedia.org/wiki/Zip_(file_format)#Compression_methods Note also that ``openpyxl`` uses ``ZIP_DEFLATED``, which seems to be the most portable if not the best compression. """ if isinstance(file, str): # it's a filename with open(file, "wb") as binaryfile: return self.write_zip(binaryfile, encoding) # recurse once with zipfile.ZipFile(file, mode="w", compression=compression) as z: # Write to ZIP. # If there are no valid task instances, there'll be no TSV; # that's OK. for filename_stem in self.get_page_names(): tsv_filename = filename_stem + ".tsv" tsv_contents = self.get_tsv_file(page_name=filename_stem) z.writestr(tsv_filename, tsv_contents.encode(encoding))
[docs] def as_zip(self, encoding: str = "utf-8") -> bytes: """ Returns the TSV collection as a ZIP file containing TSV files. Args: encoding: encoding to use when writing the TSV files """ with io.BytesIO() as memfile: self.write_zip(memfile, encoding) zip_contents = memfile.getvalue() return zip_contents
# ------------------------------------------------------------------------- # XLSX, ODS # -------------------------------------------------------------------------
[docs] def write_xlsx(self, file: Union[str, BinaryIO]) -> None: """ Write the contents in XLSX (Excel) format to a file. Args: file: filename or file-like object """ if XLSX_VIA_PYEXCEL: # use pyexcel_xlsx data = self._get_pyexcel_data(convert_for_openpyxl) pyexcel_xlsx.save_data(file, data) else: # use openpyxl # Marginal performance gain with write_only. Does not automatically # add a blank sheet wb = XLWorkbook(write_only=True) valid_name_dict = self.get_pages_with_valid_sheet_names() for page, title in valid_name_dict.items(): ws = wb.create_sheet(title=title) page.write_to_openpyxl_xlsx_worksheet(ws) wb.save(file)
[docs] def as_xlsx(self) -> bytes: """ Returns the TSV collection as an XLSX (Excel) file. """ with io.BytesIO() as memfile: self.write_xlsx(memfile) contents = memfile.getvalue() return contents
[docs] @staticmethod def get_sheet_title(page: SpreadsheetPage) -> str: r""" Returns a worksheet name for a :class:`SpreadsheetPage`. See ``openpyxl/workbook/child.py``. - Excel prohibits ``\``, ``*``, ``?``, ``:``, ``/``, ``[``, ``]`` - LibreOffice also prohibits ``'`` as first or last character but let's just replace that globally. """ title = re.sub(r"[\\*?:/\[\]']", "_", page.name) if len(title) > 31: title = f"{title[:28]}..." return title
def _get_pyexcel_data( self, converter: Callable[[Any], Any] ) -> Dict[str, List[List[Any]]]: """ Returns data in the format expected by ``pyexcel``, which is an ordered dictionary mapping sheet names to a list of rows, where each row is a list of cell values. """ data = OrderedDict() for page in self.pages: data[self.get_sheet_title(page)] = page.spreadsheetrows(converter) return data
[docs] def write_ods(self, file: Union[str, BinaryIO]) -> None: """ Writes an ODS (OpenOffice spreadsheet document) to a file. Args: file: filename or file-like object """ if ODS_VIA_PYEXCEL: # use pyexcel_ods3 data = self._get_pyexcel_data(convert_for_pyexcel_ods3) pyexcel_ods3.save_data(file, data) else: # use odswriter if isinstance(file, str): # it's a filename with open(file, "wb") as binaryfile: return self.write_ods(binaryfile) # recurse once # noinspection PyCallingNonCallable with ODSWriter(file) as odsfile: valid_name_dict = self.get_pages_with_valid_sheet_names() for page, title in valid_name_dict.items(): sheet = odsfile.new_sheet(name=title) page.write_to_odswriter_ods_worksheet(sheet)
[docs] def as_ods(self) -> bytes: """ Returns the TSV collection as an ODS (OpenOffice spreadsheet document) file. """ with io.BytesIO() as memfile: self.write_ods(memfile) contents = memfile.getvalue() return contents
[docs] def get_pages_with_valid_sheet_names(self) -> Dict[SpreadsheetPage, str]: """ Returns an ordered mapping from :class:`SpreadsheetPage` objects to their sheet names. """ name_dict = OrderedDict() for page in self.pages: name_dict[page] = self.get_sheet_title(page) self.make_sheet_names_unique(name_dict) return name_dict
[docs] @staticmethod def make_sheet_names_unique(name_dict: Dict[SpreadsheetPage, str]) -> None: """ Modifies (in place) a mapping from :class:`SpreadsheetPage` to worksheet names, such that all page names are unique. - See also :func:`avoid_duplicate_name` in ``openpxl/workbook/child.py`` - We keep the 31 character restriction """ unique_names = [] # type: List[str] for page, name in name_dict.items(): attempt = 0 while name.lower() in unique_names: attempt += 1 if attempt > 1000: # algorithm failure, better to let Excel deal with the # consequences than get stuck in a loop log.debug( f"Failed to generate a unique sheet name from {name}" ) break match = re.search(r"\d+$", name) count = 0 if match is not None: count = int(match.group()) new_suffix = str(count + 1) name = name[: -len(new_suffix)] + new_suffix name_dict[page] = name unique_names.append(name.lower())
# ------------------------------------------------------------------------- # R # -------------------------------------------------------------------------
[docs] def as_r(self) -> str: """ Returns data as an R script. This could be more sophisticated, e.g. creating factors with appropriate levels (etc.). """ now = format_datetime( get_now_localtz_pendulum(), DateFormat.ISO8601_HUMANIZED_TO_SECONDS_TZ, ) table_definition_str = "\n\n".join( page.r_data_table_definition() for page in self.pages ) script = f"""#!/usr/bin/env Rscript # R script generated by CamCOPS at {now} # ============================================================================= # Libraries # ============================================================================= library(data.table) # ============================================================================= # Data # ============================================================================= {table_definition_str} """ return script
[docs] def write_r(self, filename: str, encoding: str = "utf-8") -> None: """ Write the contents in R format to a file. Args: filename: filename or file-like object encoding: encoding to use """ with open(filename, "wt", encoding=encoding) as f: f.write(self.as_r())
def _make_benchmarking_collection( nsheets: int = 100, nrows: int = 200, ncols: int = 30, mindata: int = 0, maxdata: int = 1000000, ) -> SpreadsheetCollection: log.info( f"Creating SpreadsheetCollection with nsheets={nsheets}, " f"nrows={nrows}, ncols={ncols}..." ) coll = SpreadsheetCollection() for sheetnum in range(1, nsheets + 1): rows = [ { f"c{colnum}": str(random.randint(mindata, maxdata)) for colnum in range(1, ncols + 1) } for _ in range(1, nrows + 1) ] page = SpreadsheetPage(name=f"sheet{sheetnum}", rows=rows) coll.add_page(page) log.info("... done.") return coll
[docs]def file_size(filename: str) -> int: """ Returns a file's size in bytes. """ return os.stat(filename).st_size
[docs]def benchmark_save( xlsx_filename: str = "test.xlsx", ods_filename: str = "test.ods", tsv_zip_filename: str = "test.zip", r_filename: str = "test.R", ) -> None: """ Use with: .. code-block:: python from cardinal_pythonlib.logs import main_only_quicksetup_rootlogger from camcops_server.cc_modules.cc_spreadsheet import benchmark_save main_only_quicksetup_rootlogger() benchmark_save() Args: xlsx_filename: XLSX file to create ods_filename: ODS file to create tsv_zip_filename: TSV ZIP file to create r_filename: R script to create Problem in Nov 2019 is that ODS is extremely slow. Rough timings: - TSV ZIP: about 4.1 Mb, about 0.2 s. Good. - XLSX (via openpyxl): about 4.6 Mb, 16 seconds. - XLSX (via pyexcel_xlsx): about 4.6 Mb, 16 seconds. - ODS (via odswriter): about 53 Mb, 56 seconds. - ODS (via pyexcel_ods3): about 2.8 Mb, 29 seconds. """ coll = _make_benchmarking_collection() log.info("Writing TSV ZIP...") coll.write_zip(tsv_zip_filename) log.info(f"... done. File size {file_size(tsv_zip_filename)}") log.info("Writing XLSX...") coll.write_xlsx(xlsx_filename) log.info(f"... done. File size {file_size(xlsx_filename)}") log.info("Writing ODS...") coll.write_ods(ods_filename) log.info(f"... done. File size {file_size(ods_filename)}") log.info("Writing R...") coll.write_r(r_filename) log.info(f"... done. File size {file_size(r_filename)}")