A python module to extract all the content of an Excel document and enable calculation without Excel
Koala converts any Excel workbook into a python object that enables on the fly calculation without the need of Excel.
Koala parses an Excel workbook and creates a network of all the cells with their dependencies. It is then possible to change any value of a node and recompute all the depending cells.
Koala is available on pypi so you can just:
pip install koala2
alternatively, you can download it and install the last version from github:
git clone https://github.com/anthill/koala.git cd koala python setup.py install
Koala is still in early stages of developement and feel free to leave us issues when you encounter a problem.
The first thing you need is to convert your workbook into a graph. This operation may take some time depending on the size of your workbook (we've used koalo on workbooks containg more than 100 000 intricated formulas).
from koala.ExcelCompiler import ExcelCompiler sp = Spreadsheet("examples/basic.xlsx")
If this step fails, ensure that your Excel file is recent and in standalone mode (open it with Excel and save, it should rewrite the file and the resulting file should be three of four times heavier).
As the previous convertion can be long on big graphs, it is often useful to dump the graph to a file:
which can be relaoded later with:
sp = Spreadsheet.load('file.gzip')
You can read the values of some cells with
cell_evaluate. It will only evaluate the calculation if a parent cell has been modified with
sp.cell_set_value('Sheet1!A1', 10) sp.cell_evaluate('Sheet1!D1')
Named cells or range
If your Excel file has names defined, you can use them freely:
You can pass
ignore_sheets to ignore a list of Sheets, and
ignore_hidden to ignore all hidden cells:
sp = Spreadsheet(file, ignore_sheets = ['Sheet2'], ignore_hidden = True)
In case you have very big files, you might want to reduce the size of the output graph. Here are a few methods.
Volatiles are functions that might output a reference to Cell rather than a specific value, which impose a reevaluation everytime. Typical examples are INDEX and OFFSET.
After having created the graph, you can use
clean_pointers to fix the value of the pointers to their initial values, which reduces the graph size and decreases the evaluation times:
Warning: this implies that Cells concerned by these functions will be fixed permanently. If you evaluate a cell whose modified parents are separated by a pointer, you may encounter errors. WIP: we are working on automatic detection of the required pointers.
You can specify the outputs you need. In this case, all Cells not concerned in the calculation of these output Cell will be discarded, and your graph size wil be reduced.
sp = sp.gen_graph(inputs=['Sheet1!A1'], outputs=['Sheet1!D1', Sheet1!D2])
In this case, all Cells not impacted by inputs Cells will be discarded, and your graph size wil be reduced.
sp = sp.prune_graph()
Fix and free Cells
You might need to fix a Cell, so that its value is not reevaluated. You can do that with:
By default, all Cells on which you use
sp.cell_set_value() will be fixed.
You can free your fixed cells with:
sp.cell_free('Sheet1!D1') # frees a single Cell sp.cell_free() # frees all fixed Cells
When you free a Cell, it is automatically reevaluated.
If you need to change a Cell's formula, you can use:
sp.cell_set_formula('Sheet1!D1', 'Sheet1!A1 * 1000')
string you pass as argument needs to be written with Excel syntax.
** You will find more examples and sample excel files in the directory
To check if you have "alive pointers", i.e, pointer functions that have one of your inputs as argument, you can use:
sp.detect_alive(inputs = [...], outputs = [...])
This will also change the
Spreadsheet.pointers_to_reset list, so that only alive pointers are resetted on
Create from scratch
The graph can also be created from scratch (not by using a file).
sp_scratch = Spreadsheet() sp_scratch.cell_add('Sheet1!A1', value=1) sp_scratch.cell_add('Sheet1!A2', value=2) sp_scratch.cell_add('Sheet1!A3', formula='=SUM(Sheet1!A1, Sheet1!A2)') sp_scratch.cell_evaluate('Sheet1!A3')
This project is a "double fork" of two awesome projects:
- Pycel, a python module that generates AST graph from a workbook
- OpenPyXL, a full API able to read/write/manipulate Excel 2010 files.
The most work we did was to adapt Pycel algorithm to more complex cases that it is capable of. This ended up in modifying some core parts of the library, especially with the introduction of
As for OpenPyXL, we only took tiny bits, mainly concerning the reading part. Most of what we took from it is left unchanged in the
openpyxl folder, with references to the original scripts on BitBucket.
Release history Release notifications
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size & hash SHA256 hash help||File type||Python version||Upload date|
|koala2-0.0.35-py3-none-any.whl (87.0 kB) Copy SHA256 hash SHA256||Wheel||py3|
|koala2-0.0.35.tar.gz (69.8 kB) Copy SHA256 hash SHA256||Source||None|