
The last feature we need to implement is the ability for the user to be able to download their results in csv format. We’ve already done something very similar when we implemented the demo geometry file download, so this won’t be too taxing!

We need a new download button and it makes sense to locate this in a new tab, so we’ll add the following code to our `Parametrization()` class…

#### `app.py`

```python
tab_3 = Tab('Results Export')
# Button to download results
tab_3.section_1 = Section('Export Results')
tab_3.section_1.info = Text(info_export)
tab_3.section_1.download_btn = DownloadButton('Download Results', method='download_file')
```

Note that since we've added this to the `Parametrization()` class, the new tab will be added to the left side of our app, next to the _'Data Input'_ tab. The `info_export` text is as follows…

#### `text.py`

```python
info_export="""
## Exporting Results to CSV

The results of our analysis are reported in the graphical and tabular results panes, to the right. However, it may also be useful to export all results in a single csv file. So, we demonstrate that here. Note that we could also build a PDF report - but I'll leave that as an exercise for you!
"""
```

Note that we’ve specified a method called `download_file` to be executed when the button is clicked. We can add that to our `Controller()` class next…

#### `app.py`

```python
# Results file download
def download_file(self, params, **kwargs):
	nodes, members = processGeometryFile(params.tab_2.section_2.geometry_file)
	restraints = computeRestraintArray(params.tab_2.section_3.table_restraints)
	forces = computeForceArray(params.tab_2.section_4.table_loads)
	mbr_disp, mbr_forces, node_disp, reactions = analyseStructure(params.tab_2.section_1.E, params.tab_2.section_1.A, nodes, members, restraints, forces)
	result = exportResults(mbr_forces, node_disp, reactions, members)
	return DownloadResult(file_content=result, file_name='results.csv')
```

All of the functions that are called in `download_file()` are familiar to us by now…except one, `exportResults()`. This is the function in `dataIO.py` that will collect all of our data and compile it for download.

Once the data is compiled and passed back as `result` above, this is passed into the `DownloadResult()` component to generate the file in the user's browser, just like our previous csv file download.

Before we wrap up, we’ll review `exportResults()` in `dataIO.py`…

#### `dataIO.py`

```python
#Compile results for export
def exportResults(mbr_forces, node_disp, reactions, members):
	# Create a BytesIO object in binary mode
	output = io.BytesIO()

	# Write the data, manually encoding each row as bytes

	# Write member forces
	output.write("MEMBER FORCES, kN \n".encode('utf-8'))
	for i, mbr in enumerate(members):
		output.write(f"Member {i+1} (nodes {mbr[0]} to {mbr[1]}), {mbr_forces[i]}\n".encode('utf-8'))

	#Write nodal displacements
	output.write("\n".encode('utf-8'))
	output.write("NODAL DISPLACEMENTS, Ux (mm), Uy (mm)\n".encode('utf-8'))
	for i, d in enumerate(node_disp):
		output.write(f"Node {i+1}, {d[0]}, {d[1]} \n".encode('utf-8'))

	# Write support reactions
	output.write("\n".encode('utf-8'))
	output.write("SUPPORT REACTIONS, Rx (kN), Ry (kN)\n".encode('utf-8'))
	for i, r in enumerate(reactions):
		output.write(f"Node {r[0]}, {r[1]}, {r[2]} \n".encode('utf-8'))

	# To use the BytesIO object, we need to seek back to the start
	output.seek(0)

	return output
```

This function takes all of the calculated results as input, and starts by initialising a `BytesIO` buffer. Then, we process through each set of results, encoding each result as a `utf-8` string and adding it to the buffer. Once we’ve added all of the results to the buffer, we can return it at the bottom of the function.

We now have a functioning download button that will allow us to export our results, Fig. 1.

<Image
	src="/images/courses/using-viktor-to-build-a-shareable-truss-calculator-app/exporting-data-results/img1.png"
	alt="Results export button. | EngineeringSkills.com"
	width="1200"
	height="954"
	caption="Fig 1. Results export button."
/>

With the structure loaded as shown previously, the _'Download Results'_ button will now produce a csv file containing the following data…

```
MEMBER FORCES	 kN
Member 1 (nodes 1 to 2)	-23.75
Member 2 (nodes 2 to 3)	-23.75
Member 3 (nodes 3 to 4)	-21.25
Member 4 (nodes 4 to 5)	-21.25
Member 5 (nodes 5 to 6)	30.05
Member 6 (nodes 6 to 7)	27.95
Member 7 (nodes 7 to 8)	27.95
Member 8 (nodes 1 to 8)	33.59
Member 9 (nodes 2 to 8)	-10
Member 10 (nodes 3 to 7)	-25
Member 11 (nodes 4 to 6)	-5
Member 12 (nodes 3 to 8)	-1.77
Member 13 (nodes 3 to 6)	-5.3

NODAL DISPLACEMENTS	 Ux (mm)	 Uy (mm)
Node 1	0	0
Node 2	-0.0001	-0.00063
Node 3	-0.00019	-0.00073
Node 4	-0.00028	-0.00057
Node 5	-0.00036	0
Node 6	-5.00E-05	-0.00055
Node 7	-0.00018	-0.00058
Node 8	-0.00032	-0.00059

SUPPORT REACTIONS	 Rx (kN)	 Ry (kN)
Node 1	0	23.75
Node 5	0	21.25
```

🎉 Well done - our app is finished! You now know enough to go way beyond what we covered in this build…so get cracking and build something awesome (and don’t forget to share it)!

<Callout type="success" heading="Reporting">
	Generating a csv of our data is about the simplest type of export we could
	generate. Again, the aim here is to get you started - as an exercise, why not
	explore how to generate a pdf report of your results containing data and plots
	of the structure? You can read more about how to get started with reporting
	[here](https://docs.viktor.ai/docs/getting-started/tutorials/automatic-reporting/?prt=sean).
</Callout>

In the next section of this project, we’ll work through deploying our app and publishing it to the public.

Oh...and commit your code!


Please log in to continue