## Details:

The function get_link_preview_from_html is a common tool used in many modules, one of them documents.

When you add a link on a workspace that response with a content with a xml declaration (ex. "<!--?xml version="1.0" encoding="UTF-8"?-->"), this will raise the next ValueError:

"Unicode strings with encoding declaration are not supported. Please use bytes input or XML fragments without declaration."

At the moment that the string is being parsed as a html element, this string has been cleaned, to avoid the issue, we can only remove this xml declaration because this element is only being used to extract information about the page.

## Impacted versions:

18.0 and later

## Steps to reproduce:
1. Go to Documents App
2. Add a link in any workspace with xml declaration (ex. https://www.buffalo.jp/s3/guide/crmm/userguide/99/en/pc_index.html)

## Current behavior:
Raise Value Error "Unicode strings with encoding declaration are not supported. Please use bytes input or XML fragments without declaration."

## Expected behavior:
Link should be saved.

### Task
OPW-4675813

XLS(X) allows to store in cells date/datetimes values as date/datetime objects. Meaning, instead of having a string with the date in some format, the cell value can hold an actual date/datetime object, which are automatically converted into the `datetime.date`/`datetime.datetime` when the xls(x) file is parsed in python.

When an xls(X) file contains at the same time date values under date/datetime objects and under strings with the user format `%d/%m/%Y` (rather than the server format `%Y-%m-%d`),
the import was failing with the error
`time data '06/30/2025' does not match format '%Y-%m-%d'`

This error normally happens when the user tries to do an import with different kind of date formats under strings in the same file e.g. `06/30/2025` and `2025-07-01`
and this is understandable that Odoo doesn't know what to do in such a case.
But, if you stick to the same format, either only `%d/%m/%Y` either only '%Y-%m-%d'`, Odoo supports it.

However the case here is trickier: it's when the file contains at the same time dates under a string format, e.g. `06/30/2025` and under date objects, e.g. `datetime.date(2025, 6, 30)`. Which can happen quite easily, as Google Spreadsheet for instance tends to automatically convert the cells holding a date value into date object. And it's then easy to have a file containing both date objects and strings for date, which looks visually the same in the Google Spreadsheet interface.

In addition Google Spreadsheet tends to convert automatically only dates below the 12 of the month because of the american format. e.g.
if you set `01/06/2025`, it gets converted into a datetime object if you set `13/06/2025`, it doesn't get converted into a datetime object, the value stays as a string.

The goal of this revision is to support to have the possibility of having date values under date objects and string in a user format

The problem lied in the fact, during the xls parsing, date objects were converted into strings using the server date format. And then you could finish with data containing both the user format and the server format.

The idea is to no longer automatically convert date objects and to support having date objects in the import parsing. And when trying to guess the date/datetime format of the file, date objects are simply ignored, as they do not need to be parsed.

### Steps to reproduce:
- Accounting > Reporting > Partner Ledger
- Select multiple partners
- Select "Customer Statements" as report
- Click "Send"
- Go to the one of the partner in question an dopen the generated PDF
- All current companies show in the header and other currencies are present

### Cause:
When sending to multiple partners, the action is dispatched with to cron_threads so a new `env` is created. This `env` has a key `companies` containing all companies that exist in the DB. When the `_init_options_companies` is called for the report it adds `self.env.companies` to the option ([code](https://github.com/odoo/enterprise/blob/4ef74fd9527a28616bf81c9716d6000aa25b839e/account_reports/models/account_report.py#L1276)). So in the end all companies appear on the report.

### Solution:
Add 'forced_companies' in the options before calling the `cron`. 

The format of 'forced_companies' had to be changed to only contain ids because fields cannot be added in `send_and_print_values` as it's a JSON fields. 

opw-4635283

**Steps to reproduce:**
- Install Accounting
- In Bank journal settings, make sure that there's no Outstanding Payments account

- Create a bill for a customer
- Confirm it
- Create a second bill for the same customer
- Confirm it

- Go to the bills list
- Select the 2 created bills
- Create a payment via "Pay" button:
  * Journal: Bank
  * Payment Method: Manual Payment
  * Group Payments: [checked]
- Do not validate the payment
- Go to "Accounting / Vendors / Payments"
- Select the created payment and create a batch
- Validate the batch payment => The payment status of both bills is "In Payment"

- From Accounting dashboard, go to transactions of Bank journal (reconciliation widget)
- Create a new transaction for the previous customer with the amount of the batch payment
- In the reconciliation widget, go to "Batch Payments" tab
- Select the created batch payment
- Validate to reconcile
- Check the bills list

**Issue:**
Only one of the bills has its payment status set to "Paid", the other one staying in "In Payment" status.

**Cause:**
In the method that checks which account move lines are included in the batch payment, we loop on each linked move line and take the "min" between the remaining amount of the batch payment and the amount of the move line until the remaining amount reaches 0.
However, for vendor bills, these amounts are negative and taking the "min" amount returns the remaining amount directly, resulting in only having 1 move line set to reconcile.

opw-4595006