如何将sbt dependencyTree的纯依赖树输出保存至文件？

执行sbt dependencyTree命令时，输出开头会包含sbt版本、项目加载等前置信息，后续才是实际的依赖树内容。例如输出如下：

[info] welcome to sbt 1.8.3 (Eclipse Adoptium Java 17.0.7)
[info] loading settings for project http4s-build from build.sbt,plugins.sbt ...
[info] loading project definition from /Users/chandrapratapprajapati/sbt-local/http4s/project
[info] loading settings for project root from build.sbt ...
[info] resolving key references (63042 settings) ...
[info] set scmInfo to https://github.com/http4s/http4s
[info] set current project to http4s (in build file:/Users/chandrapratapprajapati/sbt-local/http4s/)
[info] org.http4s:http4s_2.13:0.23.19-6-b22f722-SNAPSHOT [S]
[info] org.http4s:http4s-scalafix-internal_2.13:0.23.19-6-b22f722-SNAPSHOT [S]
[info]   +-ch.epfl.scala:scalafix-core_2.13:0.10.4 [S]
[info]     +-com.geirsson:metaconfig-typesafe-config_2.13:0.10.0 [S]
[info]     | +-com.geirsson:metaconfig-core_2.13:0.10.0 [S]
[info]     | | +-com.geirsson:metaconfig-pprint_2.13:0.10.0 [S]
[info]     | | | +-com.lihaoyi:fansi_2.13:0.3.0
[info]     | | | | +-com.lihaoyi:sourcecode_2.13:0.2.7 (evicted by: 0.3.0)
[info]     | | | | +-com.lihaoyi:sourcecode_2.13:0.3.0 [S]
[info]     | | | |
[info]     | | | +-org.scala-lang:scala-compiler:2.13.10 [S]
[info]     | | | | +-io.github.java-diff-utils:java-diff-utils:4.12
[info]     | | | | +-net.java.dev.jna:jna:5.9.0

希望仅保留纯依赖树内容并保存到文件，比如：

[info] org.http4s:http4s_2.13:0.23.19-6-b22f722-SNAPSHOT [S]
[info] org.http4s:http4s-scalafix-internal_2.13:0.23.19-6-b22f722-SNAPSHOT [S]
[info]   +-ch.epfl.scala:scalafix-core_2.13:0.10.4 [S]
[info]     +-com.geirsson:metaconfig-typesafe-config_2.13:0.10.0 [S]
[info]     | +-com.geirsson:metaconfig-core_2.13:0.10.0 [S]
[info]     | | +-com.geirsson:metaconfig-pprint_2.13:0.10.0 [S]
[info]     | | | +-com.lihaoyi:fansi_2.13:0.3.0
[info]     | | | | +-com.lihaoyi:sourcecode_2.13:0.2.7 (evicted by: 0.3.0)
[info]     | | | | +-com.lihaoyi:sourcecode_2.13:0.3.0 [S]
[info]     | | | |
[info]     | | | +-org.scala-lang:scala-compiler:2.13.10 [S]
[info]     | | | | +-io.github.java-diff-utils:java-diff-utils:4.12
[info]     | | | | +-net.java.dev.jna:jna:5.9.0

实现方法 #

方法1：用grep过滤匹配依赖行 #

依赖树的行都包含组织ID:项目ID:版本号的格式，利用这个特征过滤，同时保留结构空行：

sbt dependencyTree | grep -E '^\[info\](.*:.*:.*|\s*)$' > dependency-tree.txt

这条命令会筛选出所有以[info]开头，且后续内容要么是包含至少两个冒号的依赖坐标行，要么是空行，最后将结果写入dependency-tree.txt文件。

方法2：结合sbt参数减少前置输出+过滤 #

用sbt的-batch和-no-colors参数减少启动阶段的冗余输出，再配合grep过滤：

sbt -batch -no-colors dependencyTree | grep -E '^\[info\](.*:.*:.*|\s*)$' > dependency-tree.txt

-batch关闭交互模式，减少不必要的提示；-no-colors移除终端颜色控制代码，避免干扰过滤逻辑。

方法3：用sed精准截取依赖树起始位置 #

如果能确定前置信息的结束标记（比如最后一行前置信息是set current project to...），可以用sed从该行之后开始截取：

sbt dependencyTree | sed -n '/set current project to/,$p' | grep -v 'set current project to' > dependency-tree.txt

sed -n '/set current project to/,$p'会输出从包含指定内容的行到结尾的所有内容，再用grep -v去掉那行前置信息，最终得到纯依赖树内容。

内容的提问来源于stack exchange，提问作者chandra pratap prajapati