Packaging and csv export #3
Conversation
|
Add gitignore to the root of the repo: |
src/cloupe_package_nj9/cloupe.py
Outdated
|
|
||
| # Final csv addition | ||
| with open(barcode_clusters, 'w') as barcode_clusters_file: | ||
| barcode_clusters_file.write("Barcodes,{},{}\n".format(cloupe.celltracks[0]["Name"],cloupe.celltracks[1]["Name"])) |
There was a problem hiding this comment.
This code ONLY works if there are two celltracks.
What if there are none? What if there's only one? What if there are 99?
src/cloupe_package_nj9/cloupe.py
Outdated
| barcode_clusters_file.write( | ||
| "\n".join( | ||
| [ | ||
| re.sub(r"[()' ]", "", str(pair)) # Remove parentheses, single quotes, and spaces |
There was a problem hiding this comment.
This is also fixed to only support 2 celltracks.
Should be dynamic based on the number of items in cloupe.celltracks.
re.sub(r"[()' ]", "", str(pair)) is very odd
src/cloupe_package_nj9/cloupe.py
Outdated
|
|
||
| collection = [] | ||
| # writing the barcodes to the csv files | ||
| with open(barcodes, 'w') as barcodes_file: |
There was a problem hiding this comment.
Could each one of the writes (barcodes, features, and annotations) be an individual function. If so, should they be inside Cloupe class?
Have a look at https://docs.python.org/3.9/library/csv.html for writing CSVs with quoting and all that fun.
There was a problem hiding this comment.
I had been trying various combination of csv writter with the data that we have. I was finding it difficult to dump out the csv properly even when i was using the quoting and quotechar features along with the delimiter. I was thinking if i cud stick without using csv module. Let me know ur thoughts.
Regarding moving the writes as methods. My query would be do i remove the main function entirely and the cloupe.py will remain a class?
|
Related to #2 |
…he root directory
|
Hi Martin, |
|
src/cloupe.py
Outdated
| def spatial_plot(self): | ||
| spatial_plot_attributes = vars(self) | ||
| spatial_embedding = spatial_plot_attributes["projections"]['Spatial'] | ||
| plt.scatter(x=spatial_embedding[0], y=spatial_embedding[1]) |
There was a problem hiding this comment.
we wanted to dump projections as csv, not plot them?
There was a problem hiding this comment.
If i do that wud nt the file format get changed to xls? becuz csv doesn't have that capability? correct me if I am wrong
There was a problem hiding this comment.
Only get the coordinates, no plotting, for example:
- projection_tsne.csv
- projection_Spatial.csv
Each file being:
| Barcodes | x | y | z |
|---|---|---|---|
| ACTCAT-1 | 0.3 | 0.1 | 0.0 |
| ACTCAT-1 | 0.7 | 0.1 | 0.0 |
| ... | 0.0 | 0.0 | 0.0 |
| ACTCTT-1 | 1.0 | 0.45 | 0.0 |
There was a problem hiding this comment.
Quick question y is there an z coordinate. Do I need to default set it to 0.0?
There was a problem hiding this comment.
I think axis may be different for each projection, some may have 2 others 3.
When working with a specific projection, you can get the dimensions with the first part of "Dims" that represents the number of columns you have in the table.
cols, rows = tuple(projection["Dims"])Instead of x,y,z you can use d0,d1,d2 matching each dimension in the projection.
There was a problem hiding this comment.
Done
|
Have added the projection methods. Let me know if there is anything i can do to improve or tweak the performance of those functions. Also if i need to add in methods for projections like UMAP and Fiducials? |
I have formatted the tool to take in input file via the argparse module and at the same time generate output csv files containing Barcodes, Features and Annotations. The build was meant to be used like a n executable hence was made in this format. To execute the code : --> Python3 ./cloupe.py <input_file path>
Once the script is executed it will generate .csv files in the current working directory.