Avoid propagate empty value to change_state and only set new value if it is in modification_dict

See merge request nexedi/erp5!1367

- Column Version length is extended from 10 to 255 so it can store md5sum's lenght (32 characters) that is needed by data streams created during ebulk tool ingestions used in wendelin.
- New column Size is needed to speed up ebulk tool and can generally be useful for DMS, so that we can search/sort document by size.

File is a top-level module document, like PDF, Text Document etc, so it
should have its own security definition and should not acquire local roles.

For cases where files are embedded as sub documents, we are using Embedded
File, which acquire local roles.

In "Attach document" action of erp5_ingestion, the default publication section of attach document suggested in the dialog is defined by a type based script.

We could do the same for group, so that we have a script that is easier to customize for specific cases. One use case is for accounting transactions, the group is usually the group of the entity where the document is attached.

Also, we could enable type based script in document scanner dialog, which is very similar to attach document dialog, but it was not using the scripts.

Similarly, we introduce a script to suggest a title in document scanner.

Another difference between "Attach Document" and "Scan Document" is that the later had a "publication state" field, allowing user to choose a state for the created document. We extend this to "Attach Document" as well and introduce another customizable type based method to be able to configure a default value.

See merge request nexedi/erp5!1359

They are often used to achieve surprising styling

This customizable type based method is used in "Attach Document" and
"Scan Document" dialogs, to set a default value in the dialog for the
publication state field.

This code is not used and rely on pdftk that we don't have installed in
slapos

Do not remove relative url, as they are ofter use in img tag

This allow to see the content when being displayed as a web page content

details is not always the last child

Extend "Attach Document" to have the same feature as in "Scan Document"
so that user can directly choose the state when uploading a document.

This simplifies a bit document_scanner, which can reuse the code that is
now in Base_contribute

document_scanner: introduce a customizable type based script to suggest a title for scanned document

Unlike when we attach a document, where the title comes from the filename of the attached document,
when we scan document we don't have any filename. There's a default behaviour of using the current date, so that title is not empty, but in some cases we can define a script to suggest a better title, for exemple when scanning an invoice, we can suggest using supplier name and the invoice title.

ingestion's dialog uses some type based methods for customizability, use
same type based methods here.

The default group of documents attached to accounting transactions can be
taken from the group of sections

Base_viewNewFileDialog/your_publication_section allows to easily suggest
a default publication section to the user, but the "group" was not so
easily customizable, it was just getting from document or user.

Introduce the same mechanism as for publication section, a type based
script that can be defined for each portal type.

On documents, we have an "Export" action for OOoDocument_viewConvertDialog.
When erp5_ods_style is installed, "Export to Spreadsheet" is added as a global
action. Both these actions were translated as "Exporter" in french, which is
confusing and a bit incorrect.
Use a proper, different translation for "Export to Spreadsheet"

This test was comparing lines returned in unspecified order, sort the lines
to be sure.

fix pylint warnings

See merge request nexedi/erp5!1363

"contact us" feature of erp5_web uses erp5_crm

testTemplateTool depend on the actual order of installation of business
template, so update it now that the order has changed.

configurator scripts also had to be updated, so that
erp5_ingestion_mysql_innodb_catalog is installed before erp5_web/erp5_crm
because since erp5_crm depends on ingestion, it will install
erp5_ingestion_mysql_innodb_catalog, which has new tables and SQL methods
for indexation, so it needs to be installed with `update_catalog=True`.

As we can see here, configurator uses installBusinessTemplateListFromRepository
to install business templates one by one (in the same order that they are
added with addConfigurationItem), and pass update_catalog only for
business templates which add new tables/indexation methods and needs a
catalog update after install, but what happened is that we install
erp5_web with update_catalog=False, which will install
erp5_ingestion_mysql_innodb_catalog, also with update_catalog=False so
the activities to install business templates fail with errors like:

    ProgrammingError: (1146, "Table 'erp5_test_*.email' doesn't exist")

Since 459093f3 (ERP5Form: implement a basic checkConsistency for
ProxyField, 2021-01-26) we can detect inconsistentencies in ProxyFields
internal data structures. These are the changes after using
checkConsistency(fixit=True) for erp5_web

This relation field is used to select the web pages that will be displayed,
by using the standard ERP5 document management feature of selecting the
best applicable page with the same reference, but this list did not show
the reference of the web page