bitcoin_block_parser/
utxos.rs

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
//! Contains [`UtxoParser`] for tracking input amounts and output statuses in [`UtxoBlock`].

use crate::blocks::{BlockParser, ParserIterator, ParserOptions, Pipeline};
use anyhow::{bail, Result};
use bitcoin::block::Header;
use bitcoin::hashes::Hash;
use bitcoin::{Amount, Block, OutPoint, Transaction, TxIn, TxOut, Txid};
use dashmap::DashMap;
use log::info;
use rand::prelude::SmallRng;
use rand::{Error, RngCore, SeedableRng};
use scalable_cuckoo_filter::{DefaultHasher, ScalableCuckooFilter, ScalableCuckooFilterBuilder};
use std::fs;
use std::fs::File;
use std::io::{BufReader, BufWriter};
use std::iter::Zip;
use std::slice::Iter;
use std::sync::{Arc, Mutex};

/// A block that has been parsed tracking input amounts and output status
#[derive(Clone, Eq, PartialEq, Debug)]
pub struct UtxoBlock {
    /// The block header
    pub header: Header,
    /// List of transactions contained in the block
    pub txdata: Vec<UtxoTransaction>,
}

impl UtxoBlock {
    /// Construct from a bitcoin [`Block`].
    fn new(block: Block) -> Self {
        Self {
            header: block.header,
            txdata: block.txdata.into_iter().map(UtxoTransaction::new).collect(),
        }
    }

    /// Convert back into a [`bitcoin::Block`].
    pub fn to_block(self) -> Block {
        Block {
            header: self.header,
            txdata: self.txdata.into_iter().map(|tx| tx.transaction).collect(),
        }
    }
}

/// A transaction that has been parsed tracking input amounts and output status
#[derive(Clone, Eq, PartialEq, Debug)]
pub struct UtxoTransaction {
    /// Underlying bitcoin transaction [`Transaction`]
    pub transaction: Transaction,
    /// Precomputed [`Txid`]
    pub txid: Txid,
    /// Tracks the input amounts in-order of inputs
    inputs: Vec<Amount>,
    /// Tracks the output statuses in-order of outputs
    outputs: Vec<OutputStatus>,
}

impl UtxoTransaction {
    /// Construct from a bitcoin [`Transaction`].
    fn new(transaction: Transaction) -> UtxoTransaction {
        Self {
            txid: transaction.compute_txid(),
            transaction,
            inputs: vec![],
            outputs: vec![],
        }
    }

    /// Returns the [`TxIn`] of the transaction zipped with the input amounts.
    pub fn input(&self) -> Zip<Iter<'_, TxIn>, Iter<'_, Amount>> {
        self.transaction.input.iter().zip(self.inputs.iter())
    }

    /// Returns the [`TxOut`] of the transaction zipped with the output [`OutputStatus`].
    pub fn output(&self) -> Zip<Iter<'_, TxOut>, Iter<'_, OutputStatus>> {
        self.transaction.output.iter().zip(self.outputs.iter())
    }
}

/// Status of the [`TxOut`] within the transaction graph.
#[derive(Clone, Debug, Eq, PartialEq, Copy)]
pub enum OutputStatus {
    /// The output was spent in a later block.
    Spent,
    /// The output was never spent in any later block (it is a UTXO).
    Unspent,
    /// The status of the output is unknown (only if [`UtxoParser::load_filter`] was not called).
    Unknown,
}

type ShortOutPoints = (Vec<ShortOutPoint>, Vec<ShortOutPoint>);
type ShortOutPointFilter = ScalableCuckooFilter<ShortOutPoint, DefaultHasher, FastRng>;

/// Multithreaded parser that returns a [`ParserIterator`] of [`UtxoBlock`].
/// * Tracks the [`Amount`] for every [`TxIn`].
/// * Tracks the [`OutputStatus`] for every [`TxOut`] if [`UtxoParser::load_filter`] is called.
///
/// # Examples
/// Computing the largest mining fee requires knowing the input amounts of every transaction.
/// Call [`UtxoParser::parse`] to get a [`UtxoBlock`] that tracks input amounts.
/// ```no_run
/// use std::cmp::max;
/// use bitcoin::Amount;
/// use bitcoin_block_parser::utxos::*;
///
/// let parser = UtxoParser::new("/home/user/.bitcoin/blocks/").unwrap();
/// let fees = parser.parse().map_parallel(|block| {
///     let mut max_mining_fee = Amount::ZERO;
///     for tx in block.txdata.into_iter() {
///         // For every transaction sum up the input and output amounts
///         let inputs: Amount = tx.input().map(|(_, amount)| *amount).sum();
///         let outputs: Amount = tx.output().map(|(out, _)| out.value).sum();
///         if !tx.transaction.is_coinbase() {
///             // Subtract outputs amount from inputs amount to get the fee
///             max_mining_fee = max(inputs - outputs, max_mining_fee);
///         }
///     }
///     max_mining_fee
/// });
/// println!("Maximum mining fee: {}", fees.max().unwrap());
/// ```
///
/// Computing the largest UTXO requires knowing the [`OutputStatus`] to determine whether a
/// [`TxOut`] was spent or unspent.  Call [`UtxoParser::load_or_create_filter`] to track the output
/// status.
///
/// Although this takes longer to run the first time it also lowers the memory usage.
/// ```no_run
/// use std::cmp::max;
/// use bitcoin::Amount;
/// use bitcoin_block_parser::utxos::*;
///
/// let parser = UtxoParser::new("/home/user/.bitcoin/blocks/").unwrap();
/// let blocks = parser.load_or_create_filter("filter.bin").unwrap().parse();
/// let amounts = blocks.map_parallel(|block| {
///     let mut max_unspent_tx = Amount::ZERO;
///     for tx in block.txdata.into_iter() {
///         for (output, status) in tx.output() {
///             if status == &OutputStatus::Unspent {
///                 max_unspent_tx = max(output.value, max_unspent_tx);
///             }
///         }
///     }
///     max_unspent_tx
/// });
/// println!("Maximum unspent output: {}", amounts.max().unwrap());
/// ```
pub struct UtxoParser {
    /// Filter that contains all unspent transaction outpoints.
    filter: Option<ShortOutPointFilter>,
    /// Underlying parser for parsing the blocks.
    parser: BlockParser,
    /// Used to allocate the initial capacity of shared state.
    estimated_utxos: usize,
}

impl UtxoParser {
    /// Creates a new parser given the `blocks` directory where the `*.blk` files are located.
    ///
    /// - Returns an `Err` if unable to parse the `blk` files.
    /// - You can [specify the blocks directory](https://en.bitcoin.it/wiki/Data_directory) when
    ///   running `bitcoind`.
    pub fn new(blocks_dir: &str) -> Result<Self> {
        Self::new_with_opts(blocks_dir, ParserOptions::default())
    }

    /// Creates a parser with custom [`ParserOptions`].
    pub fn new_with_opts(blocks_dir: &str, options: ParserOptions) -> Result<Self> {
        Ok(Self {
            filter: None,
            parser: BlockParser::new_with_opts(blocks_dir, options)?,
            estimated_utxos: 300_000_000,
        })
    }

    /// Set the estimated amount of UTXOs in the range of blocks you are parsing.
    ///
    /// Used to lower the memory usage of shared state objects.
    pub fn estimated_utxos(mut self, estimated_utxos: usize) -> Self {
        self.estimated_utxos = estimated_utxos;
        self
    }

    /// Parse the blocks into an iterator of [`UtxoBlock`].
    pub fn parse(self) -> ParserIterator<UtxoBlock> {
        // if using a filter we can save memory by reducing the initial hashmap capacity
        let hashmap_capacity = if self.filter.is_some() {
            self.estimated_utxos / 10
        } else {
            self.estimated_utxos
        };
        let pipeline = UtxoPipeline::new(self.filter, hashmap_capacity);
        self.parser
            .parse(UtxoBlock::new)
            .ordered()
            .pipeline(&pipeline)
    }

    /// Set the height of the last block to parse.
    ///
    /// Parsing always starts at the genesis block in order to track the transaction graph properly.
    pub fn block_range_end(mut self, end: usize) -> Self {
        self.parser = self.parser.block_range(0, end);
        self
    }

    /// Loads a `filter_file` or creates a new one if it doesn't exist.
    pub fn load_or_create_filter(self, filter_file: &str) -> Result<Self> {
        if !fs::exists(filter_file)? {
            self.create_filter(filter_file)?.load_filter(filter_file)
        } else {
            self.load_filter(filter_file)
        }
    }

    /// Loads a `filter_file` or returns `Err` if it doesn't exist.
    pub fn load_filter(mut self, filter_file: &str) -> Result<Self> {
        if !fs::exists(filter_file)? {
            bail!("Filter file '{}' doesn't exist", filter_file);
        }
        let reader = BufReader::new(File::open(filter_file)?);
        let filter = bincode::deserialize_from(reader)?;

        self.filter = Some(filter);
        Ok(self)
    }

    /// Creates a new `filter_file`.
    pub fn create_filter(self, filter_file: &str) -> Result<Self> {
        info!("Creating '{}'", filter_file);
        let filter = UtxoFilter::new(self.estimated_utxos);
        self.parser
            .parse(UtxoFilter::outpoints)
            .ordered()
            .map(&|outpoints| filter.update(outpoints))
            .for_each(|_| {});

        let filter = Arc::try_unwrap(filter.filter).expect("Arc still referenced");
        let mut filter = Mutex::into_inner(filter)?;
        filter.shrink_to_fit();
        let writer = BufWriter::new(File::create(filter_file)?);
        bincode::serialize_into(writer, &filter)?;
        Ok(self)
    }
}

/// Contains the filter data that tracks all unspent outputs in a memory-efficient manner.
#[derive(Clone)]
struct UtxoFilter {
    filter: Arc<Mutex<ShortOutPointFilter>>,
}

impl UtxoFilter {
    /// Construct with an initial `filter_capacity`.
    fn new(filter_capacity: usize) -> UtxoFilter {
        Self {
            filter: Arc::new(Mutex::new(
                ScalableCuckooFilterBuilder::default()
                    .initial_capacity(filter_capacity)
                    .false_positive_probability(0.000_000_000_001)
                    .rng(FastRng::default())
                    .finish(),
            )),
        }
    }

    /// Returns [`ShortOutPoint`] for all inputs and outputs.
    fn outpoints(block: Block) -> ShortOutPoints {
        let mut inputs = vec![];
        let mut outputs = vec![];
        for tx in block.txdata.iter() {
            let txid = tx.compute_txid();
            for input in &tx.input {
                inputs.push(ShortOutPoint::from_outpoint(&input.previous_output));
            }

            for (index, _) in tx.output.iter().enumerate() {
                outputs.push(ShortOutPoint::new(index, &txid));
            }
        }
        (inputs, outputs)
    }

    /// Given the results of `outpoints()` update the filter.
    pub fn update(&self, outpoints: ShortOutPoints) {
        let mut filter = self.filter.lock().expect("Lock poisoned");
        let (inputs, outputs) = outpoints;
        for outpoint in outputs {
            // insert outpoints for every output
            filter.insert(&outpoint);
        }
        for input in inputs {
            // remove outpoints that are spent in a subsequent transaction
            filter.remove(&input);
        }
    }
}

/// Pipeline for multithreaded tracking of the input amounts and output statuses.
#[derive(Clone, Default)]
struct UtxoPipeline {
    /// Optional filter containing all unspent outpoints.
    filter: Option<Arc<ShortOutPointFilter>>,
    /// Tracks the amounts for every input.
    amounts: Arc<DashMap<ShortOutPoint, Amount>>,
}

impl UtxoPipeline {
    /// Construct a new pipeline with an optional `filter` and initial `hashmap_capacity`.
    fn new(filter: Option<ShortOutPointFilter>, hashmap_capacity: usize) -> Self {
        Self {
            filter: filter.map(Arc::new),
            amounts: Arc::new(DashMap::with_capacity(hashmap_capacity)),
        }
    }

    /// Returns the [`OutputStatus`] of an outpoint, returning [`OutputStatus::Unknown`] if running
    /// without a filter.
    fn status(&self, outpoint: &ShortOutPoint) -> OutputStatus {
        match &self.filter {
            None => OutputStatus::Unknown,
            Some(filter) if filter.contains(outpoint) => OutputStatus::Unspent,
            _ => OutputStatus::Spent,
        }
    }
}

impl Pipeline<UtxoBlock, UtxoBlock, UtxoBlock> for UtxoPipeline {
    fn first(&self, mut block: UtxoBlock) -> UtxoBlock {
        for tx in &mut block.txdata {
            for (index, output) in tx.transaction.output.iter().enumerate() {
                let outpoint = ShortOutPoint::new(index, &tx.txid);
                let status = self.status(&outpoint);
                // if an outpoint is unspent we don't need to track it (saving memory)
                if status != OutputStatus::Unspent {
                    self.amounts.insert(outpoint, output.value);
                }
                tx.outputs.push(status);
            }
        }
        block
    }

    fn second(&self, mut block: UtxoBlock) -> UtxoBlock {
        for tx in &mut block.txdata {
            for input in tx.transaction.input.iter() {
                if tx.transaction.is_coinbase() {
                    // coinbase transactions will not have a previous input
                    tx.inputs.push(Amount::ZERO);
                } else {
                    let outpoint = ShortOutPoint::from_outpoint(&input.previous_output);
                    let (_, value) = self.amounts.remove(&outpoint).expect("Missing outpoint");
                    tx.inputs.push(value);
                }
            }
        }
        block
    }
}

/// Shortened [`OutPoint`] to save memory (14 bytes instead of 36 bytes)
///
/// - 2 bytes represent far more than the maximum tx outputs (2^16)
/// - 12 byte subset of the txid is unlikely to generate collisions even with 1 billion txs (~6.3e-12)
#[derive(Eq, PartialEq, Hash, Debug, Clone)]
struct ShortOutPoint(pub Vec<u8>);
impl ShortOutPoint {
    /// Shorten an existing [`OutPoint`].
    fn from_outpoint(outpoint: &OutPoint) -> ShortOutPoint {
        Self::new(outpoint.vout as usize, &outpoint.txid)
    }

    /// Create a new [`ShortOutPoint`] given its transaction id and index.
    fn new(vout: usize, txid: &Txid) -> ShortOutPoint {
        let mut bytes = vec![];
        bytes.extend_from_slice(&vout.to_le_bytes()[0..2]);
        bytes.extend_from_slice(&txid.as_byte_array()[0..12]);
        ShortOutPoint(bytes)
    }
}

/// Wrapper for [`SmallRng`] since it doesn't implement [`Default`] required to deserialize.
#[derive(Debug)]
struct FastRng(SmallRng);
impl Default for FastRng {
    fn default() -> Self {
        Self(SmallRng::seed_from_u64(0x2c76c58e13b3a812))
    }
}
impl RngCore for FastRng {
    fn next_u32(&mut self) -> u32 {
        self.0.next_u32()
    }

    fn next_u64(&mut self) -> u64 {
        self.0.next_u64()
    }

    fn fill_bytes(&mut self, dest: &mut [u8]) {
        self.0.fill_bytes(dest)
    }

    fn try_fill_bytes(&mut self, dest: &mut [u8]) -> std::result::Result<(), Error> {
        self.0.try_fill_bytes(dest)
    }
}