Some improvement ideas

[thluffy-gif]

Hello everyone,

Congratulations on this project! I've been looking for a self-hosted solution to manage my wine cellar for a long time, and Wine Cellar is exactly what I needed. The fact that it's open-source, Docker-compatible, and supports multiple users is perfect for sharing with my bro.

After exploring the application, here are some features that would be really useful for wine enthusiasts like me:

Top A ranking based on user ratings in the cellar:

=> Top 10/20 highest-rated wines
=> Filters by type, region, year
=> Statistics: average rating, number of tastings

Imaybe with : A Django view with aggregation of existing ratings, pagination, and filters reusing current models. Could be added to navigation without changing the architecture.

Financial Tracking and Valuation

=> Visual comparison of purchase price vs current value
=> Cellar ROI calculation
=> Price evolution over time

maybe with : Extend the Bottle model with purchase_price and current_market_price fields. View with Chart.js for graphs. Compatible with existing architecture.

Simple Opening Status

=> Mark bottle as "opened/closed"
=> Opening date tracking
=> Filter inventory by status

maybe with BooleanField is_opened + DateTimeField opened_date on Bottle model. Quick action in list views. Simple migration without breaking changes.

Enhanced Photo Management

=> uploads (label)

maybe with : Local or S3 storage via django-storages (already compatible).

OIDC Authentication

For those with self-hosted infrastructure (Authelia/Keycloak):

=> OpenID Connect support
=> SSO with other services

Integration approach: mozilla-django-oidc package as additional auth backend. Configuration via environment variables. Works alongside existing Django authentication without conflicts.

MariaDB Compatibility

Does the project work well with MariaDB? With Django ORM it should be transparent, but for a large collection (several hundred bottles), it's reassuring to have confirmation about performance

[goapunk]

Hi @thluffy-gif
glad you like the app! And happy to get some feedback.

Top A ranking based on user ratings in the cellar:

=> Top 10/20 highest-rated wines
=> Filters by type, region, year
=> Statistics: average rating, number of tastings

It's a good idea and in theory not very hard to build. However, currently the models are designed to make sure other users can't learn about what wines you have in your cellar (or you drank in the past etc.). This makes implementing this a bit harder, as you'd need to match identical wines between users somehow (probably doable without too much effort). Additionally, some kind of user setting to have a user consent their wines are shown in a statistic accessible to everyone would be necessary. I will create an issue and look into it.

Financial Tracking and Valuation

=> Visual comparison of purchase price vs current value
=> Cellar ROI calculation
=> Price evolution over time

maybe with : Extend the Bottle model with purchase_price and current_market_price fields. View with Chart.js for graphs. >Compatible with existing architecture.

To be honest I'm not into investing into bottles (yet?) nor do I have an extensive collection of bottles, so I'm not really familiar with the requirements for such a feature. I think a first step could be to extend the bottle/stock mechanism with an optional price field and then show an average (lowest,highest) price for the wine (calculated from the bottle prices), as prices might fluctuate (even outside investments). A field for the current market price would probably make more sense on the wine model (as it's not tied to a specific bottle but to the wine itself) ? I will create issues for the idea but have to admit it will be not on top of the priority list. I'm of course happy to accept a PR for it :)

Simple Opening Status

=> Mark bottle as "opened/closed"
=> Opening date tracking
=> Filter inventory by status

maybe with BooleanField is_opened + DateTimeField opened_date on Bottle model. Quick action in list views. Simple migration >without breaking changes.

I think that's a great idea as currently you can only track when a bottle/vintage should be used up but not open bottles in general. I guess this is more relevant for red and fortified wines, as whites and roses generally don't have a long fridge life?

Enhanced Photo Management

=> uploads (label)

maybe with : Local or S3 storage via django-storages (already compatible).

Yes, having the option to upload more pictures and tag them (e.g. front, back, ...) is a good idea. Do you think some predefined tags are enough or should you be able to add custom ones?

OIDC Authentication

For those with self-hosted infrastructure (Authelia/Keycloak):

=> OpenID Connect support
=> SSO with other services

That's definitely on the list! I'd probably go with django-allauth as I have more experience with it and it should support most self-hosted SSOs.

MariaDB Compatibility

Does the project work well with MariaDB? With Django ORM it should be transparent, but for a large collection (several hundred >bottles), it's reassuring to have confirmation about performance

I haven't tested this project with MariaDB but I don't see a reason why not, it should be pretty equal to postgresql. For a collection with less than a few thousand bottles I'd expect it to even work well enough with sqlite3. Swapping postgresql with mariadb should be fairly trivial within the docker-compose file as well.

[thluffy-gif]

Of course, it is always enjoyable to discuss interesting projects with fellow enthusiasts.
I will take the time to provide a bit more detail in my messages. My previous message was a list of ideas that came to mind, some more useful than others.

Personally, I don’t really invest in wine, and there are already specialized sites doing an excellent job, with open or paid APIs (maybe for a future integration). My initial idea was more focused on a personal cellar. It is always nice to have a valuation of one’s cellar and to view some charts, even if they only reflect the initial purchase price, for example:
=> Total value of consumed wine
=> Total value of available wine

(This is more of a fun feature than an essential one from my point of view, and I agree with you on that.)

Indeed, I hadn’t considered the multi-user aspect. A robust idea to match wines could be to store them in a shared database when a new wine record is created, which would then enable a preselection if another user wants to add the same wine.
Absolutely, I see this feature as mostly useful for red wine, especially since there are convenient systems with CO2 injection to preserve an opened bottle and be able to pour a glass later, which is very handy for a personal cellar.

Regarding tags, that’s a good question. I think an interesting first approach would be to allow saving the label. Personally, I can’t imagine managing 50 different tags: what really matters for a bottle is having photos of the front and back to see the label clearly. I struggle to see the usefulness of multiple tags, but I would be curious to know how you would use several?

I will try to run some tests with MariaDB soon and keep you informed.

While writing this response, I thought that a demo instance of the project could be a nice addition to showcase it on the GitHub page (maybe hosted on a free VM like an e2-micro from Google or Oracle).

[goapunk]

Thanks :)

Total value of consumed wine
Total value of available wine
Ah, that makes sense, and should be fairly easy!

Indeed, I hadn’t considered the multi-user aspect. A robust idea to match wines could be to store them in a shared database when a new wine record is created, which would then enable a preselection if another user wants to add the same wine.

Initially I wanted add a feature which allows sharing the wines with a central server and build an open source / open data of all wines, which would also allow us to search for a wine before having to enter it manually. I guess we could start locally with that as you described it.

Regarding tags, that’s a good question. I think an interesting first approach would be to allow saving the label. Personally, I can’t imagine managing 50 different tags: what really matters for a bottle is having photos of the front and back to see the label clearly. I struggle to see the usefulness of multiple tags, but I would be curious to know how you would use several?

I guess for now tags like "Front", "Back", "Label" would probably be enough ? And if someone wants to upload other pictures they can let us know here.

While writing this response, I thought that a demo instance of the project could be a nice addition to showcase it on the GitHub page (maybe hosted on a free VM like an e2-micro from Google or Oracle).

A demo instance would be cool! I don't have experience with any of the "free" boxes though, so will have to see which one is viable and really free.