From 0cd97d584ec437ca11f4b88bbf0e549135fb0ea7 Mon Sep 17 00:00:00 2001
From: Wouter Deconinck <wdconinc@gmail.com>
Date: Thu, 22 Aug 2024 15:43:08 -0500
Subject: [PATCH] RPackage: parse description for incomplete dependencies

---
 lib/spack/spack/build_systems/r.py | 60 ++++++++++++++++++++++++++++++
 1 file changed, 60 insertions(+)

diff --git a/lib/spack/spack/build_systems/r.py b/lib/spack/spack/build_systems/r.py
index 87b1e104aeb..b7e7e41df9e 100644
--- a/lib/spack/spack/build_systems/r.py
+++ b/lib/spack/spack/build_systems/r.py
@@ -1,12 +1,19 @@
 # Copyright Spack Project Developers. See COPYRIGHT file for details.
 #
 # SPDX-License-Identifier: (Apache-2.0 OR MIT)
+import re
 from typing import Optional, Tuple
 
+import llnl.util.filesystem as fs
+import llnl.util.lang as lang
+import llnl.util.tty as tty
+import spack.deptypes as dt
 from llnl.util.filesystem import mkdirp
 from llnl.util.lang import ClassProperty, classproperty
 
+from spack.dependency import Dependency
 from spack.directives import extends
+from spack.spec import Spec
 
 from .generic import GenericBuilder, Package
 
@@ -38,6 +45,59 @@ def install(self, pkg, spec, prefix):
         """Installs an R package."""
         mkdirp(pkg.module.r_lib_dir)
 
+        try:
+            # TODO: use a more sustainable dcf parser
+            import pycran
+
+            # Read DESCRIPTION file with dependency information
+            # https://r-pkgs.org/description.html
+            r_deps = []
+            with open(fs.join_path(self.stage.source_path, 'DESCRIPTION')) as file:
+                for desc in pycran.parse(file.read()):
+                    r_deps.extend([d.strip() for d in desc.get("Imports", None).split(",")])
+                    r_deps.extend([d.strip() for d in desc.get("Depends", None).split(",")])
+
+            # Convert to spack dependencies format for comparison
+            deps = {}
+            for r_dep in r_deps:
+                p = re.search(r"^[\w_-]+", r_dep)  # first word, incl. underscore or dash
+                v = re.search("(?<=[(]).*(?=[)])", r_dep)  # everything between parentheses
+                # require valid package
+                assert(p, f"Unable to find package name in {r_dep}")
+                r_spec = f"r-{p[0].strip().lower()}" if p[0].lower() != "r" else "r"
+                # allow minimum or pinned versions
+                if v:
+                    v = re.sub(r">=\s([\d.-]+)", r"@\1:", v[0])  # >=
+                    v = re.sub(r"==\s([\d.-]+)", r"@\1", v)  # ==
+                    deps[r_spec] = Dependency(pkg, Spec(r_spec + v), dt.BUILD | dt.RUN)
+                else:
+                    deps[r_spec] = Dependency(pkg, Spec(r_spec), dt.BUILD | dt.RUN)
+
+            # Retrieve dependencies for current spack package and version
+            spack_dependencies = []
+            for when, dep in pkg.dependencies.items():
+                if spec.satisfies(when):
+                    spack_dependencies.append(dep)
+            merged_dependencies = {}
+            for dep in spack_dependencies:
+                for n, d in dep.items():
+                    if d in merged_dependencies:
+                        merged_dependencies[n].merge(d)
+                    else:
+                        merged_dependencies[n] = d
+
+            # For each R dependency, ensure Spack dependency is at least as strong
+            for dep in sorted(deps.keys()):
+                if dep in merged_dependencies:
+                    if not merged_dependencies[dep].spec.satisfies(deps[dep].spec):
+                        tty.debug(f'    depends_on("{deps[dep].spec}", when="@{pkg.version}:", type=("build", "run"))')
+                else:
+                    tty.debug(f'    depends_on("{deps[dep].spec}", when="@{pkg.version}:", type=("build", "run"))')
+
+        except ImportError:
+            tty.debug("R package dependency verification requires pycran")
+            pass
+
         config_args = self.configure_args()
         config_vars = self.configure_vars()