Some Python-based tools for exploring the Nano block-lattice and LMDB-based database that the Nano wallet and node software at  use. Nano (https://nano.org) is a cryptocurrency that aims to be feeless, instant and high throughput. Note that this project has no relation to the official Nano project.
Included is a tool to convert (a subset of) stored blocks to a SQLite database for easy querying with SQL. Also includes an explorer similar to the one on nano.org, mostly for testing, but also usable to browse the data.
In case you find this software useful, donations are welcome at xrb_1mycqeczobsiyerohkmeeyt7ehyyfjyz5h4hi53ffb6p5qjrefzfcrpc454t
Paul Melis (firstname.lastname@example.org)
- All Python scripts are developed for Python 3.x. They might work for 2.x as well, but this isn't tested.
- In similar fashion the main development platform is Linux, although Windows and macOS should work as well. If you find issues on the non-Linux platforms please submit an issue report at https://github.com/paulmelis/nano-tools/issues
- On 32-bit systems, or when using a 32-bit version of Python on a 64-bit system, it is not possible to read the Nano LMDB database when its size is >2GB. This limitation has to do with the memory-mapping used by LMDB. In practice, a fully synced node/wallet currently uses more than 4GB, so using the tools below isn't possible on these systems.
The interesting scripts are:
- Low-level tool to inspect the contents of the LMDB database used by the Nano wallet/node software.
- Main script to convert the LMDB-based Nano/RaiBlocks database to a SQLite database
- Close the official Nano wallet/node software, so nothing else is
accessing the LMDB database (which by default is at
$ ./conv2sqlite.py convert
This will create a SQLite database called
nano.dbin the current directory. This step will take a couple of minutes for a fully synced Nano wallet/node, also depending on the speed of the disk being written to.
- You should now have a SQLite database file
- Close the official Nano wallet/node software, so nothing else is accessing the LMDB database (which by default is at
- Note: the SQLite database is by default written in the current directory.
You can change the output file with the
- If you have enough free memory (say 4-8 GBs) you can
generate the SQLite database on a ram-disk, such as
/dev/shmon Linux, for faster generation and improved query performance. Copy it to a persistent disk later if needed.
- A Python module that provides an object-oriented API to the SQLite database
conv2sqlite.py. This allows easy querying and navigation of blocks, accounts and relations between them. The explorer uses this API.
- A Python module that provides an object-oriented API to the SQLite database created by
- A web-based account and block explorer similar to https://nano.org/en/explore/. It lacks certain features and is available mostly to inspect the generated SQLite DB, but can also be used to browse and explore the Nano block-lattice.
$ ./explorer.py nano.db
- Browse to http://localhost:7777/known_accounts
- Utility module containing some routines to work with native Nano values, such as accounts, balances and amounts.
- lmdb (conv2sqlite.py, dump_wallet_db.py)
- numpy (dump_wallet_db.py only)
- click (conv2sqlite.py only)
- Flask (explorer.py only)
Different versions of these packages will probably work. Development is done using APSW 3.21.0, lmdb 0.93, numpy 1.14.1, click 6.7, flask 0.12.2.
See the LICENSE file in this source distribution.
The included Bootstrap, jQuery and Popper sources (included under the 'static' directory) have their own license.
What operating systems will this work on?
- In principle it should work everywhere Python and the required dependencies are available. In practice, it is developed and tested on Linux. So with other operating systems there might be some things not fully working (yet).
Why is the code based on Python 3.x?
- The main reason for using Python 3 was that it comes with an implementation of the blake2b hashing algorithm used in Nano, which is used in the routines that work on accounts.
Why convert to a separate SQLite database instead of working directly on the LMDB database that the Nano wallet/node software uses?
- It is safer to work on a separate copy of the wallet/node database, in case anything goes wrong when working with the data.
- SQL is a very powerful language for querying. Also, LMDB does not offer automatic indexing of the data, leading to potentially slow queries.
- An LMDB database only contains key-value pairs, that need to be decoded into separate fields. In contrast, a SQL table contains columns, one per field, for easier querying.
- It is easier and safer to add custom data in a separate table in the SQLite database than it is in the LMDB database.
Is it safe to have the official Nano wallet/node running while converting the LMDB database to SQLite?
- This has not been tested by the author of this software. It is recommended that you NOT have the wallet/node running while doing the conversion, regardless of what the LMDB docs say about concurrent access. Also, these scripts might not handle unexpected changes to the LMDB database, while reading it, very well.
Can I update the database after letting the wallet/node receive new blocks?
- This is currently not implemented. The only way to update is to rebuild the SQLite database.
Why use APSW instead of the built-in sqlite3 module?
- APSW is an excellent library, aimed at working with SQLite from Python, while sqlite3 follows the DB-API which isn't SQLite-specific and somewhat quirky to use. The sqlite3 module (depending on your Python version) also has issues with transactions in certain scenarios. All in all, APSW is much more pleasant to work with. As it is available in many distros as a package or can be installed using pip adding it to your system shouldn't be too hard.