Merge pull request #718 from betatim/early-r

[MRG] Install R packages before copying repo contents

Author: minrk

repo2docker/buildpacks/base.py

@@ -90,7 +90,7 @@
 
 {% if build_script_files -%}
 # If scripts required during build are present, copy them
-{% for src, dst in build_script_files.items() %}
+{% for src, dst in build_script_files|dictsort %}
 COPY {{ src }} {{ dst }}
 {% endfor -%}
 {% endif -%}
@@ -125,7 +125,14 @@
 # of the repository but don't access any files in the repository. By executing
 # them before copying the repository itself we can cache these steps. For
 # example installing APT packages.
-{% for sd in pre_assemble_script_directives -%}
+{% if preassemble_script_files -%}
+# If scripts required during build are present, copy them
+{% for src, dst in preassemble_script_files|dictsort %}
+COPY src/{{ src }} ${REPO_DIR}/{{ dst }}
+{% endfor -%}
+{% endif -%}
+
+{% for sd in preassemble_script_directives -%}
 {{ sd }}
 {% endfor %}
 
@@ -144,7 +151,7 @@
 # Container image Labels!
 # Put these at the end, since we don't want to rebuild everything
 # when these change! Did I mention I hate Dockerfile cache semantics?
-{% for k, v in labels.items() %}
+{% for k, v in labels|dictsort %}
 LABEL {{k}}="{{v}}"
 {%- endfor %}
 
@@ -379,6 +386,19 @@ def get_build_scripts(self):
 
         return []
 
+    def get_preassemble_script_files(self):
+        """
+        Dict of files to be copied to the container image for use in preassembly.
+
+        This is copied before the `build_scripts`, `preassemble_scripts` and
+        `assemble_scripts` are run, so can be executed from either of them.
+
+        It's a dictionary where the key is the source file path in the
+        repository and the value is the destination file path inside the
+        repository in the container.
+        """
+        return {}
+
     def get_preassemble_scripts(self):
         """
         Ordered list of shell snippets to build an image for this repository.
@@ -499,13 +519,13 @@ def render(self):
                 "RUN {}".format(textwrap.dedent(script.strip("\n")))
             )
 
-        pre_assemble_script_directives = []
+        preassemble_script_directives = []
         last_user = "root"
         for user, script in self.get_preassemble_scripts():
             if last_user != user:
-                pre_assemble_script_directives.append("USER {}".format(user))
+                preassemble_script_directives.append("USER {}".format(user))
                 last_user = user
-            pre_assemble_script_directives.append(
+            preassemble_script_directives.append(
                 "RUN {}".format(textwrap.dedent(script.strip("\n")))
             )
 
@@ -516,7 +536,8 @@ def render(self):
             env=self.get_env(),
             labels=self.get_labels(),
             build_script_directives=build_script_directives,
-            pre_assemble_script_directives=pre_assemble_script_directives,
+            preassemble_script_files=self.get_preassemble_script_files(),
+            preassemble_script_directives=preassemble_script_directives,
             assemble_script_directives=assemble_script_directives,
             build_script_files=self.get_build_script_files(),
             base_packages=sorted(self.get_base_packages()),

repo2docker/buildpacks/r.py

@@ -291,15 +291,52 @@ def get_build_scripts(self):
 
         return super().get_build_scripts() + scripts
 
-    def get_assemble_scripts(self):
-        """
-        Return series of build-steps specific to this repository.
+    def get_preassemble_script_files(self):
+        files = {}
+        installR_path = self.binder_path("install.R")
+        if os.path.exists(installR_path):
+            files[installR_path] = installR_path
+
+        return files
+
+    def get_preassemble_scripts(self):
+        """Install contents of install.R
+
+        Attempt to execute `install.R` before copying the contents of the
+        repository. We speculate that most of the time we do not need access.
+        In case this fails we re-run it after copying the repository contents.
+
+        The advantage of executing it before copying is that minor edits to the
+        repository content will not trigger a re-install making things faster.
         """
+        scripts = []
+
+        installR_path = self.binder_path("install.R")
+        if os.path.exists(installR_path):
+            scripts += [
+                (
+                    "${NB_USER}",
+                    "Rscript %s && touch /tmp/.preassembled || true" % installR_path,
+                )
+            ]
+
+        return super().get_preassemble_scripts() + scripts
+
+    def get_assemble_scripts(self):
+        """Install the dependencies of or the repository itself"""
         assemble_scripts = super().get_assemble_scripts()
 
         installR_path = self.binder_path("install.R")
         if os.path.exists(installR_path):
-            assemble_scripts += [("${NB_USER}", "Rscript %s" % installR_path)]
+            assemble_scripts += [
+                (
+                    "${NB_USER}",
+                    # only run install.R if the pre-assembly failed
+                    "if [ ! -f /tmp/.preassembled ]; then Rscript {}; fi".format(
+                        installR_path
+                    ),
+                )
+            ]
 
         description_R = "DESCRIPTION"
         if not self.binder_dir and os.path.exists(description_R):