machine: support instance size & startup script config (iterative#6537)

* machine: install dvc via startup_script by default

* machine: support configuring instance type and startup script

* accept startup_script as a config relpath

* use pushd/popd in default install script

dvc/config.py

@@ -263,6 +263,7 @@ def _map_dirs(conf, func):
                     "key_path": func,
                 }
             },
+            "machine": {str: {"startup_script": func}},
         }
         return Schema(dirs_schema, extra=ALLOW_EXTRA)(conf)
 

dvc/config_schema.py

@@ -229,6 +229,8 @@ class RelPath(str):
     "machine": {
         str: {
             "cloud": All(Lower, Choices("aws", "azure")),
+            "instance_type": Lower,
+            "startup_script": str,
         },
     },
     # section for experimental features

dvc/machine/__init__.py

@@ -25,6 +25,18 @@
 
 RESERVED_NAMES = {"local", "localhost"}
 
+DEFAULT_STARTUP_SCRIPT = """#!/bin/bash
+sudo apt-get update
+sudo apt-get install --yes python3 python3-pip
+python3 -m pip install --upgrade pip
+
+pushd /etc/apt/sources.list.d
+sudo wget https://dvc.org/deb/dvc.list
+sudo apt-get update
+sudo apt-get install --yes dvc
+popd
+"""
+
 
 def validate_name(name: str):
     from dvc.exceptions import InvalidArgumentError
@@ -168,6 +180,12 @@ def _get_backend(self, cloud: str) -> "BaseMachineBackend":
     def create(self, name: Optional[str]):
         """Create and start the specified machine instance."""
         config, backend = self.get_config_and_backend(name)
+        if "startup_script" in config:
+            with open(config["startup_script"]) as fobj:
+                startup_script = fobj.read()
+        else:
+            startup_script = DEFAULT_STARTUP_SCRIPT
+        config["startup_script"] = startup_script
         return backend.create(**config)
 
     def destroy(self, name: Optional[str]):