An independent service of transaction graph

[Keith-CY]

utxo 模型 交易的拓扑是一个网图
可视化出来应该还是比较有意思的

All cells on-chain are linked by input -> output in transactions, so they construct a graph that shows the history of the whole chain/live cells.

It would be very intuitive and impressive for data analysis.

The service can be designed standalone and provides data to CKB explorer UI

[homura]

In my imagination, it may look like this

---

At the present, there doesn't seem to be a way to fetch which transaction consumed a particular OutPoint via RPC directly

---

A close idea is to use the get_transactions to probe which transactions have consumed the Script of the OutPoint, but there are some problems with this way

• the scripts of the OutPoint must not be popular on the chain, otherwise the probing process will become very long

[Keith-CY]

In my imagination, it may look like this

At the present, there doesn't seem to be a way to fetch which transaction consumed a particular OutPoint via RPC directly

A close idea is to use the get_transactions to probe which transactions have consumed the Script of the OutPoint, but there are some problems with this way

• the scripts of the OutPoint must not be popular on the chain, otherwise the probing process will become very long

Please have a look at this proposal @Danie0918 @Sven-TBD

[Danie0918]

In my imagination, it may look like this
At the present, there doesn't seem to be a way to fetch which transaction consumed a particular OutPoint via RPC directly
A close idea is to use the get_transactions to probe which transactions have consumed the Script of the OutPoint, but there are some problems with this way

• the scripts of the OutPoint must not be popular on the chain, otherwise the probing process will become very long

Please have a look at this proposal @Danie0918 @Sven-TBD

This is a useful and interesting feature and we will start moving forward with its implementation.

[Sven-TBD]

https://vs0cjf.axshare.com/?id=j6zrfc&p=_354_an_independent_service_of_transaction_graph&g=1
@Keith-CY

[Keith-CY]

vs0cjf.axshare.com/?id=j6zrfc&p=_354_an_independent_service_of_transaction_graph&g=1 @Keith-CY

cc @homura

[homura]

The diagram is cool and inspiring, but I guess it might be affected by the above one. It is based on the view of a single cell. However, it's not easy to define an entry cell, and it's difficult, maybe impossible, to define a one-to-one relationship between cells. Therefore, I propose changing the view from single-cell-based to timeline-based

Next, I'll talk about what each of these elements represents.

---

A cellbase transaction example

---

This is a chained transaction exam. The dashed lines linked to the transactions are clickable and link to the corresponding block area.

---

This is an example of the spent cellbase transaction. Also, the dashed line is clickable and links to the corresponding block area.

[Keith-CY]

The diagram is cool and inspiring, but I guess it might be affected by the above one. It is based on the view of a single cell. However, it's not easy to define an entry cell, and it's difficult, maybe impossible, to define a one-to-one relationship between cells. Therefore, I propose changing the view from single-cell-based to timeline-based

Next, I'll talk about what each of these elements represents.

A cellbase transaction example This is a chained transaction exam. The dashed lines linked to the transactions are clickable and link to the corresponding block area. This is an example of the spent cellbase transaction. Also, the dashed line is clickable and links to the corresponding block area.

Is it necessary to order the transactions by block [n]? Transactions are ordered by their nature(linked list). The graph may be hard to read if a transaction is linked to several transactions that are far from each other from the block number dimension because they may not be linear. For instance

Block                   Block + N            Block + N + 1000            Block + N + 10000000
Transaction ---------> Transaction
     ├-----------------------------------------> Transaction
     └-----------------------------------------------------------------------> Transaction

[homura]

Is it necessary to order the transactions by block [n]?

Block                   Block + N            Block + N + 1000            >Block + N + 10000000
Transaction ---------> Transaction
    ├-----------------------------------------> Transaction
    └--------------------------------------------------------->--------------> Transaction

No, but it isn't easy to design how to

• choose a cell as an entry
• show cells between Block + N and Block + N + 1000

The new one I proposed is not a fix, but another new way that will make the design more intuitive

The graph may be hard to read if a transaction is linked to several transactions that are far from each other from the block number dimension because they may not be linear

Exactly, this is the reason to keep an OutPoint with a dashed line in the Tx3, and users can click the OutPoint with a left-slide animation to locate the corresponding block area

[homura]

A PoC can be previewed here https://ckb-tx-illustration.vercel.app/

• The node size is controlled by log10(capacity) instead of capacity because the amount of capacity varies greatly from cell to cell
• Source code is here https://github.com/homura/ckb-tx-illustration

[yanguoyu]

I have already implemented a demo using Cytoscape. Cytoscape is convenient for constructing tree diagrams. I will continue to research how to add nodes dynamically.

[FrederLu]

https://pudge.explorer.nervos.org/transaction/0x8ab56f4f70cac7c23aac75bc81abb403f799befc16490528f342b3837fc82824

https://pudge.explorer.nervos.org/transaction/0xbf824fe159e97d65af0776c7ad9261fc98f5b648f411edf9a4bab20bf63a43e2

It was found that the fifth data in UTXO graph was incomplete. @yanguoyu

[Sven-TBD]

https://pudge.explorer.nervos.org/transaction/0x8ab56f4f70cac7c23aac75bc81abb403f799befc16490528f342b3837fc82824

https://pudge.explorer.nervos.org/transaction/0xbf824fe159e97d65af0776c7ad9261fc98f5b648f411edf9a4bab20bf63a43e2

It was found that the fifth data in UTXO graph was incomplete. @yanguoyu

I checked with Guoyu like last week, and I told him 4 cells should be displayed here by default, though he said it could be 5, never mind, let's make it 4, so we could mark this bug as fixed @FrederLu

[FrederLu]

https://pudge.explorer.nervos.org/transaction/0x8ab56f4f70cac7c23aac75bc81abb403f799befc16490528f342b3837fc82824
https://pudge.explorer.nervos.org/transaction/0xbf824fe159e97d65af0776c7ad9261fc98f5b648f411edf9a4bab20bf63a43e2
It was found that the fifth data in UTXO graph was incomplete. @yanguoyu

I checked with Guoyu like last week, and I told him 4 cells should be displayed here by default, though he said it could be 5, never mind, let's make it 4, so we could mark this bug as fixed @FrederLu

OK, I got it then.