class Thumb(db.Entity): _table_ = 'thumbnails' Id = orm.PrimaryKey(str) Image = orm.Optional(bytes) changed the definition and now works fine (when everything fails, read the manual) :)

Hi Nick! Thank you for this! I can review it on Saturday

Found some interesting stuff after spending some time looking at the decompiler, running against our codebase, and got a couple of tests that capture the issues. It looks like python 3.12 will generate different bytecode depending on whitespace, and I was able to write a couple of tests that will pass on 3.10, but fail on 3.12 (pony misses part of the AST when decompiling the bytecode) Was wondering if anyone more famililar with the decompiler would be able to take a look and take over or offer some pointers. https://github.com/ponyorm/pony/pull/741/files#diff-5d718a04b57f46be1e1d57d8aafeea6fe8223ca90ce99c1eaeda219d5837b6f4R257-R258

🍽 Кулинарный лайфхак! 🍽 🎄Розмарин — это ароматная и универсальная специя, которая идеально дополняет множество блюд. Вот список блюд, к которым он подходит лучше всего: 🍖 Мясные блюда: жаркое из баранины или ягненка, свинина, говядина. 🐟 Рыба и морепродукты: запеченный лосось или другая жирная рыба, креветки на гриле, рыбные супы или бульоны. 🥔 Овощи и гарниры: запеченный картофель (дольки или целый картофель), тыква или кабачки, грибы. ♥️ #КулинарныеСоветы #Розмарин #ИдеиДляКухни

Есть места на УДАЛЕНКУ с хорошим доходом. — Занятость 1-2 часа в день Пишите в ЛС "+" за деталями!

Есть места на УДАЛЕНКУ с хорошим доходом. — Занятость 1-2 часа в день Пишите в ЛС "+" за деталями!

Есть места на УДАЛЕНКУ с хорошим доходом. — Занятость 1-2 часа в день Пишите в ЛС "+" за деталями!

good morning, have some issue with entities serialization. say that have these two entities class File(db.Entity): ... location = orm.Optional('Location') class Location(db.Entity): ... file = orm.Optional('File') that relationship may be represented as CREATE TABLE "File" ( "id" INTEGER PRIMARY KEY AUTOINCREMENT, ... "location" INTEGER REFERENCES "Location" ("id") ON DELETE SET NULL ) now when doing [obj.to_dict() from obj in orm.select(x for x in db.Location)] it generates one query to retrieve the Location objects SELECT "id", ... FROM "Location" WHERE "id" IN (?, ?, ?, ...) but then it generates n queries (one for Location) to get the relationship like SELECT "id", ..., "location" FROM "File" WHERE "location" = ? LIMIT 2 wondering how to avoid this n+1 issue

def location_summary(data): location, file_id = data return ( location.to_dict(exclude=('file',)) | {'file': file_id} ) def location_select(): return [ location_summary(obj) for obj in orm.left_join( (x, x.file.id) for x in db.Location) ] this would work...

Hello. Try [obj.to_dict() from obj in orm.select(x for x in db.Location).prefetch(db.File)]

still see the queries, just now with a different id order SELECT "id", ..., "location" FROM "File" WHERE "location" = ? LIMIT 2 and then _after_ those queries see SELECT "id", ..., "location" FROM "File" WHERE "id" IN (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?) [93, 495, 1020, 2073, 2307, 2588, 2611, 2785, 2984, 5003, 5817, 8201, 8435, 11331] where those ids are the ones that do have relationship

by the way, there is no issue for (n,m) or (1,n) relationships

Hello, any update on python3.13 support?

Sorry, was overloaded at work. I hope to be able to do it tomorrow

Hello Alexander, have you got a chance to take a look to determine how much effort should go to new release? thanks a lot

Working with bytecode is always unpredictable

However, I have a question how many of you migrated to 3.13 already? On my work I still cant pass 3.10 cause of dependencies..(

I am on 3.12, only blocker currently is pony, but thats just me :)

I am on 3.11 blocked by the pop jump if none issue to move to 3.12

However, I have a question how many of you migrated to 3.13 already? On my work I still cant pass 3.10 cause of dependencies..(

Migrated to 3.13, but we are a startup with not much legacy

However, I have a question how many of you migrated to 3.13 already? On my work I still cant pass 3.10 cause of dependencies..(

Migrated to 3.13. It's the default interpreter on the latest Fedora.

good morning, ¿should use on_connect() decorator to load extensions? for example @db.on_connect(provider='sqlite') def sqlite_case_sensitivity(db, connection): connection.load_extension(sqlite_icu.extension_path().replace('.so',''))

Hi, how can I load an complete sql file through pony in my database? E.g. I have a sql dump as backup and want to refill my database

I use db.execute() now and read line by line

You sure you need Pony for this? It would be much easier to apply dump directly

My unit tests use pony and I have all the infrastructure to connect to the dB server etc in pony

My unit tests use pony and I have all the infrastructure to connect to the dB server etc in pony

Its okay but if you have sql dump its usually a bunch of inserts. Which is easier to apply with db cli

Hello @metaprogrammer, any news on py3.13? thanks in advance

Wasnt there a PR for 3.13 support?

Hello @metaprogrammer, any news on py3.13? thanks in advance

Sorry, I was overloaded these weeks at work... I will try the next weekend.

@luckydonald hello. If I'm not mistaken you were the one who made migrations approach. How did it end? Did you make it all by yourself or you used our migrations branch?

I did not use migration branch. I use regular ponyorm. The thing is at https://github.com/luckydonald/pony_up

@luckydonald hello. If I'm not mistaken you were the one who made migrations approach. How did it end? Did you make it all by yourself or you used our migrations branch?

^

Last time I used it, it worked well.

Aight thanks!

Please "select" function not works fine when pass lambda in it This happens only with python 3.13 In python 3.12 works fine

Please "select" function not works fine when pass lambda in it This happens only with python 3.13 In python 3.12 works fine

Pony does not support 3.13 yet

Although progress started

Hello, @metaprogrammer any updates on py 3.13? thx

It's me who implements this. For now I'm kinda stuck at some bug(

@metaprogrammer I added stubs with type annotations. They work fine with pyright/pylance. With mypy there are some issues and I don't think they're solvable without fixing at least 1-2 issues in mypy. Do you want a PR?

Here is the repo: https://github.com/henri-hulski/pony/tree/stubs

In the test_types branch I applied the annotations directly to the code to test integrity.

Thank you! I'll look into it

Hi, does pony serialize the access to a sqlite dB for multiple requests? For example if I have a website and multiple users using it to store data?

Does Pony have strawberr-gql support

good morning, i'm seeing a weird bug with inheritance. the value of the classtype is not retrieved from db from pony import orm from pprint import pprint def create(db): class Location(db.Entity): connection = orm.Set('Connection') @classmethod def serialize(cls, ids): return { x[0].id: x[0].custom_summary(*x[1:]) for x in orm.left_join((x, x.connection) for x in cls if x.id in ids) } def custom_summary(self, connection): origen = { 'classtype': connection.classtype, 'id': connection.id, } return self.to_dict() | { 'origen_model': origen, } def to_dict(self): return super().to_dict(exclude=['connection']) class Connection(db.Entity): location = orm.Optional('Location') class Antena(db.Connection): pass class Wireless(db.Connection): pass db = orm.Database() create(db) db.bind(provider='sqlite', filename=':memory:', create_db=create) db.generate_mapping(check_tables=True, create_tables=True) with orm.db_session: antenna = db.Antena() location = db.Location() location.connection.add(antenna) orm.set_sql_debug(debug=True, show_values=True) with orm.db_session: location = db.Location[1] loc_serial = location.serialize([location.id]) antenna_data = loc_serial[1]['origen_model'] pprint(antenna_data) antenna = db.Connection[1] pprint(antenna.to_dict())

GET CONNECTION FROM THE LOCAL POOL SWITCH TO AUTOCOMMIT MODE SELECT "id" FROM "Location" WHERE "id" = ? [1] SELECT DISTINCT "x"."id", "connection"."id" FROM "Location" "x" LEFT JOIN "Connection" "connection" ON "x"."id" = "connection"."location" WHERE "x"."id" IN (?) [1] {'classtype': 'Connection', 'id': 1} SELECT "id", "location", "classtype" FROM "Connection" WHERE "id" = ? [1] {'classtype': 'Antena', 'id': 1, 'location': 1} RELEASE CONNECTION as you can see, the first output says 'classtype': 'Connection' where it should be 'Antena'

i'm on version 0.7.19

found a fix class Location: def custom_summary(self, connection): origen = { 'foo': connection.foo, #¡important!, before classtype access 'classtype': connection.classtype, 'id': connection.id, } ... class Connection(db.Entity): location = orm.Optional('Location') foo = orm.Optional(str) # just to have something to access this way it does make a select for Connection table to retrieve the data

Hello, how does it seem with py 3.13 support, any news?

Hello, how does it seem with py 3.13 support, any news?

I guess for now it's up to open-source community to contribute. Neither me nor Alexander K could find time to finish this task.

alrighty, thanks for honest response

Are there plans to support import/export of designer models in PonyORM online editor (at least for the commercial version)?

Hi Vlad! Recently, the development of PonyORM and its online editor stalled due to a lack of time for core developers. I believe it will finally change in the near future, as I will have the opportunity to return to active development. Initially, I'll plan to focus on PonyORM itself, so I don't expect any significant improvements to the online editor soon. Still, I'm interested in hearing what type of import/export you have in mind. Can you explain it in more detail?

Hi Vlad! Recently, the development of PonyORM and its online editor stalled due to a lack of time for core developers. I believe it will finally change in the near future, as I will have the opportunity to return to active development. Initially, I'll plan to focus on PonyORM itself, so I don't expect any significant improvements to the online editor soon. Still, I'm interested in hearing what type of import/export you have in mind. Can you explain it in more detail?

When our startup finally becomes a unicorn, I will definitely donate to the development of Pony ORM!

Hi Vlad! Recently, the development of PonyORM and its online editor stalled due to a lack of time for core developers. I believe it will finally change in the near future, as I will have the opportunity to return to active development. Initially, I'll plan to focus on PonyORM itself, so I don't expect any significant improvements to the online editor soon. Still, I'm interested in hearing what type of import/export you have in mind. Can you explain it in more detail?

We use the design model for maintaining our ERD, and generating the ORM code. Basically, the design model is, kind of, the ground truth for us. The generated ORM code is stored in the project's source control. So we would like to store the design model there as well, so that both could be synchronized. We know that snapshots can be created in the editor, but this means that extra effort needs to be introduced to keep the version of ORM code and the snapshot name in sync somehow ( after all, this snapshot is kept only in PonyORM editor ) The situation is getting worse if we take into account different branches of the project. For example, we need to provide a hot-fix in some old branch by changing some aspect of the model, and would like to re-generate the ORM code - which design model snapshot should we use? Should we reflect the branch name and revision in the name of the snapshot? Seems unnecessarily complicated.

We use the design model for maintaining our ERD, and generating the ORM code. Basically, the design model is, kind of, the ground truth for us. The generated ORM code is stored in the project's source control. So we would like to store the design model there as well, so that both could be synchronized. We know that snapshots can be created in the editor, but this means that extra effort needs to be introduced to keep the version of ORM code and the snapshot name in sync somehow ( after all, this snapshot is kept only in PonyORM editor ) The situation is getting worse if we take into account different branches of the project. For example, we need to provide a hot-fix in some old branch by changing some aspect of the model, and would like to re-generate the ORM code - which design model snapshot should we use? Should we reflect the branch name and revision in the name of the snapshot? Seems unnecessarily complicated.

Thanks for the explanation, now I understand what you mean. Some time ago, we had already started working on this feature, but then paused our progress. Hope we can resume it. By the way, what DBMS do you use?

Thanks for the explanation, now I understand what you mean. Some time ago, we had already started working on this feature, but then paused our progress. Hope we can resume it. By the way, what DBMS do you use?

PostgreSQL

Thanks for the explanation, now I understand what you mean. Some time ago, we had already started working on this feature, but then paused our progress. Hope we can resume it. By the way, what DBMS do you use?

Same here