Convert Irex Notes of new firmware 2.0 BETA

[GodzillaNL]

I have created a little application which can convert the .note version to XPS or PDF documents so they can be read on a desktop computer.

The .NET framework 3.5 SP1 is required to run, just extract the RAR file and run. To associate the application with all .note files, you must run it as administrator once.

Please let me hear if you find any problems or suggestions.

Release 4:
- Added test option to merge scribbles on PDF files to PDF files, please let me know if you find any problems!
- Added device type option in options menu, depending on this option the note files will be correctly sized to fill the page thanks to Mr. X for the note file.

Release 3:
- Added the option to export to PDF instead of XPS. Change in options menu.

Release 2:
- New version where the stroke information is correctly read from the blob field.

Release 1:
- Initial release

[Matthijs]

Cool.
The sources for liberscribble are mostly unchanged from v1.0 to v2.0 so it's safe to use them as a starting point to parse the scribble data BLOB.

[GodzillaNL]

Thanks for the information. The only error I think that is causing some missed lines is that the pen size is not retrieved from the blob data, resulting in missed lines (still almost not noticable). The binary data of the blob must contain this somewhere. This is the only thing I haven't figured out yet.
All data is now of the same pen size set default in the application.

[Mackx]

Hi GodzillaNL,

I attached a file that I use for xournal to read the annotations, maybe it is helpfull for you.

Regards,
Marcel.

[CoolDragon]

Is the note format same as the one used for PDF scribbles?

[GodzillaNL]

I don't know, I haven't tested this, maybe it is the same. Renaming the metadata.db to *.note will maybe work.

I have created this application because I like the new fast notepad application in the new firmware. I was really missing some export function to the desktop. If the notepad application could support some sort of paper templates it would be superb. Like line paper, 1 cm block paper, etc.

[CoolDragon]

Well, the metadata.db file is a database file (MySQL Lite?). I am not sure if the access interface to this database file is open or not, but it certainly contains the scribbling data of each PDF file. And if I am correct, the PDF scribble and the note share the same base code, so it is possible to get the scribble data out from the metadata.db file, and merge them into the PDF file, actually that's how the DR800 and DR1000 are currently doing.

[Mr. X]

metadata.db is a sqlite3 database. All the code for it is open, see libermetadb. The Scribbles are stored as "blobs" (generic binary data) in the "annotations" table:

CREATE TABLE annotations (
  annotation_id    INTEGER PRIMARY KEY AUTOINCREMENT NOT NULL,
  annotation_type  TEXT NOT NULL,
  file_id          INTEGER NOT NULL,
  layer            TEXT NOT NULL,
  file_position    INTEGER NOT NULL,
  title            TEXT,
  start_anchor     TEXT,
  end_anchor       TEXT,
  data             BLOB
);

Presumably the blob is what you need to put into a .note file after finding the right annotation record.

[GodzillaNL]

Originally Posted by Mr. X

metadata.db is a sqlite3 database. ...

I know, I read this information. The only problem that still needs to solved is some intepretation errors while reading through the BLOB data. When I have time for this I will have a look, take a look at the problem and see if there are any problems or make suggestions.

[Mackx]

Originally Posted by Mr. X

Presumably the blob is what you need to put into a .note file after finding the right annotation record.

A .note file is also an sqlite database, that contains, among others, the annotations-table that contains the blob with annotation data. Both metadata.db- and .note-files contain the same tables that are relevant for annotations so the annotation specific code can be used either on a metadata.db file or a .note file I guess.

[Mackx]

Originally Posted by GodzillaNL

The only problem that still needs to solved is some intepretation errors while reading through the BLOB data. When I have time for this I will have a look, take a look at the problem and see if there are any problems or make suggestions.

I had to do simular things when I added the annotation-merge-functionality to xournal. So if it is a specific problem, you can send me a PM, maybe I can help.

[GodzillaNL]

Originally Posted by Mackx

I had to do simular things when I added the annotation-merge-functionality to xournal. So if it is a specific problem, you can send me a PM, maybe I can help.

I looks like the merge functionality is working perfectly. It is already in the version it attached in this thread. The only problem is in reading the BLOB data for generating the XPS file. Some errors may occur because I had to change all C code to C# code.

When this is fixed, I think of adding some more functions like moving pages in the .note file.

Can you explain how I should interpretate the stroke information in the BLOB field?

[Mackx]

Originally Posted by GodzillaNL

Can you explain how I should interpretate the stroke information in the BLOB field?

Did you have time to look at the file I send a few post back. I copied some structures from the erscribble-library that can be laid down onto the blob-data and wrote some functions to parse the blob data.

From my notes the interpretation of a very simple blob structure, hope this helps with reading the code and the structure of the blob:
0100 version
00000000 orientation
0000000000000000 page_pixel
0000 page_r
0000 page_g
0000 page_b
00000000 ul_x
00000000 ul_y
D7040000 lr_x -> 04d7
DA060000 lr_y -> 06da
02000000 num_strokes -> 2

STOKE:
03000000 color
A6838942 zoom <-- float
00000000 layer
00000000 shape
03000000 width -> 3
02000000 points -> 2

POINTS:
2e000000 x
3a000000 y
00000000 pressure
2e000000 x
3a000000 y
2C5D1600 pressure <-- not sure what to do with this

STROKE:
03000000 color
A6838942 zoom <-- float
00000000 layer
00000000 shape
03000000 width
05000000 points

POINTS:
30000000 x
44000000 y
00000000 pressure
2f000000 x
45000000 y
58C94000 pressure
30000000 x
45000000 y
58C94000 pressure
31000000 x
45000000 y
58C94000 pressure
2f000000 x
45000000 y
2C5D1600 pressure

The code also contains the following comment:
// Best results gives the following formula where:
// points from file are scaled using the zoom factor of the stroke_device
// then they are scaled from 'input' DPI (160) to 'output' DPI (72)
// for some reason a factor of (1-1/16) is needed to scale all to
// the correct resolution (maybe related to margins?)
So I also had to 'play' a little bit to get it correct.

Hope this helps.

[GodzillaNL]

Thanks, that really helps. I can already see that i made a mistake in parsing the stroke information. I hope I can solve this, this week.

[GodzillaNL]

Attached in first post is a new version, now all information is read perfectly from the BLOB field resulting is correct XPS documents.