Source code for planemo.conda_verify.recipe

import os
import re
from os.path import (

import yaml

from planemo.conda_verify.const import (
from planemo.conda_verify.utils import (

sel_pat = re.compile(r"(.+?)\s*\[(.+)\]$")
name_pat = re.compile(r"[a-z0-9_][a-z0-9_\-\.]*$")
version_pat = re.compile(r"[\w\.]+$")
url_pat = re.compile(r"(ftp|http(s)?)://")

[docs]class RecipeError(Exception): pass
[docs]def ns_cfg(cfg): plat = cfg["plat"] py = cfg["PY"] np = cfg["NPY"] for x in py, np: assert isinstance(x, int), x return dict( nomkl=False, debug=False, linux=plat.startswith("linux-"), linux32=bool(plat == "linux-32"), linux64=bool(plat == "linux-64"), armv7l=False, arm=False, ppc64le=False, osx=plat.startswith("osx-"), unix=plat.startswith(("linux-", "osx-")), win=plat.startswith("win-"), win32=bool(plat == "win-32"), win64=bool(plat == "win-64"), x86=plat.endswith(("-32", "-64")), x86_64=plat.endswith("-64"), py=py, py3k=bool(30 <= py < 40), py2k=bool(20 <= py < 30), py26=bool(py == 26), py27=bool(py == 27), py33=bool(py == 33), py34=bool(py == 34), py35=bool(py == 35), np=np, )
[docs]def select_lines(data, namespace): lines = [] for line in data.splitlines(): line = line.rstrip() m = sel_pat.match(line) if m: if PEDANTIC: x = # error on comment, unless the whole line is a comment if "#" in x and not x.startswith("#"): raise RecipeError("found commented selector: %s" % line) cond = if eval(cond, namespace, {}): lines.append( continue lines.append(line) return "\n".join(lines) + "\n"
@memoized def yamlize(data): res = yaml.safe_load(data) # ensure the result is a dict if res is None: res = {} return res
[docs]def parse(data, cfg): if cfg is not None: data = select_lines(data, ns_cfg(cfg)) # ensure we create new object, because yamlize is memoized return dict(yamlize(data))
[docs]def get_field(meta, field, default=None): section, key = field.split("/") submeta = meta.get(section) if submeta is None: submeta = {} res = submeta.get(key) if res is None: res = default return res
[docs]def check_name(name): if name: name = str(name) else: raise RecipeError("package name missing") if not name_pat.match(name) or name.endswith((".", "-", "_")): raise RecipeError("invalid package name '%s'" % name) seq = get_bad_seq(name) if seq: raise RecipeError("'%s' is not allowed in " "package name: '%s'" % (seq, name))
[docs]def check_version(ver): if ver: ver = str(ver) else: raise RecipeError("package version missing") if not version_pat.match(ver): raise RecipeError("invalid version '%s'" % ver) if ver.startswith(("_", ".")) or ver.endswith(("_", ".")): raise RecipeError("version cannot start or end with '_' or '.': %s" % ver) seq = get_bad_seq(ver) if seq: raise RecipeError(f"'{seq}' not allowed in version '{ver}'")
[docs]def check_build_number(bn): if not (isinstance(bn, int) and bn >= 0): raise RecipeError("build/number '%s' (not a positive interger)" % bn)
[docs]def check_requirements(meta): for req in get_field(meta, "requirements/run", []): name = req.split()[0] if not name_pat.match(name): raise RecipeError("invalid run requirement name '%s'" % name)
[docs]def check_license_family(meta): if not PEDANTIC: return lf = get_field(meta, "about/license_family", get_field(meta, "about/license")) if lf not in LICENSE_FAMILIES: print( """\ Error: license_family is invalid: %s Note that about/license_family falls back to about/license. Allowed license families are:""" % lf ) for x in LICENSE_FAMILIES: print(" - %s" % x) raise RecipeError("wrong license family")
[docs]def check_url(url): if not url_pat.match(url): raise RecipeError("not a valid URL: %s" % url)
[docs]def check_about(meta): summary = get_field(meta, "about/summary") if summary and len(summary) > 80: msg = "summary exceeds 80 characters" if PEDANTIC: raise RecipeError(msg) else: print("Warning: %s" % msg) for field in ("about/home", "about/dev_url", "about/doc_url", "about/license_url"): url = get_field(meta, field) if url: check_url(url) check_license_family(meta)
hash_pat = { "md5": re.compile(r"[a-f0-9]{32}$"), "sha1": re.compile(r"[a-f0-9]{40}$"), "sha256": re.compile(r"[a-f0-9]{64}$"), }
[docs]def check_source(meta): src = meta.get("source") if not src: return fn = src.get("fn") if fn: for ht in "md5", "sha1", "sha256": hexgigest = src.get(ht) if hexgigest and not hash_pat[ht].match(hexgigest): raise RecipeError("invalid hash: %s" % hexgigest) url = src.get("url") if url: check_url(url) git_url = src.get("git_url") if git_url and (src.get("git_tag") and src.get("git_branch")): raise RecipeError("cannot specify both git_branch and git_tag")
[docs]def validate_meta(meta): for section in meta: if PEDANTIC and section not in FIELDS: raise RecipeError("Unknown section: %s" % section) submeta = meta.get(section) if submeta is None: submeta = {} for key in submeta: if PEDANTIC and key not in FIELDS[section]: raise RecipeError("in section %r: unknown key %r" % (section, key)) check_name(get_field(meta, "package/name")) check_version(get_field(meta, "package/version")) check_build_number(get_field(meta, "build/number", 0)) check_requirements(meta) check_about(meta) check_source(meta)
[docs]def validate_files(recipe_dir, meta): for field in "test/files", "source/patches": flst = get_field(meta, field) if not flst: continue for fn in flst: if PEDANTIC and fn.startswith(".."): raise RecipeError("path outsite recipe: %s" % fn) path = join(recipe_dir, fn) if isfile(path): continue raise RecipeError("no such file '%s'" % path)
[docs]def iter_cfgs(): for py in 27, 34, 35: for plat in "linux-64", "linux-32", "osx-64", "win-32", "win-64": yield dict(plat=plat, PY=py, NPY=111)
[docs]def dir_size(dir_path): return sum(sum(getsize(join(root, fn)) for fn in files) for root, unused_dirs, files in os.walk(dir_path))
[docs]def check_dir_content(recipe_dir): disallowed_extensions = ( ".tar", ".tar.gz", ".tar.bz2", ".tar.xz", ".so", ".dylib", ".la", ".a", ".dll", ".pyd", ) for root, unused_dirs, files in os.walk(recipe_dir): for fn in files: fn_lower = fn.lower() if fn_lower.endswith(disallowed_extensions): if PEDANTIC: raise RecipeError("found: %s" % fn) else: print("Warning: found: %s" % fn) path = join(root, fn) # only allow small archives for testing if PEDANTIC and fn_lower.endswith((".bz2", ".gz")) and getsize(path) > 512: raise RecipeError("found: %s (too large)" % fn) if basename(recipe_dir) == "icu": return # check total size od recipe directory (recursively) kb_size = dir_size(recipe_dir) / 1024 kb_limit = 512 if PEDANTIC and kb_size > kb_limit: raise RecipeError("recipe too large: %d KB (limit %d KB)" % (kb_size, kb_limit)) if PEDANTIC: try: with open(join(recipe_dir, ""), "rb") as fi: data = if data and not data.decode("utf-8").startswith(("#!/bin/bash\n", "#!/bin/sh\n")): raise RecipeError("not a bash script:") except OSError: pass
[docs]def render_jinja2(recipe_dir): import jinja2 loaders = [jinja2.FileSystemLoader(recipe_dir)] env = jinja2.Environment(loader=jinja2.ChoiceLoader(loaders)) template = env.get_or_select_template("meta.yaml") return template.render(environment=env)
[docs]def validate_recipe(recipe_dir, pedantic=True): global PEDANTIC PEDANTIC = bool(pedantic) meta_path = join(recipe_dir, "meta.yaml") with open(meta_path, "rb") as fi: data = if PEDANTIC and not all_ascii(data): raise RecipeError("non-ASCII in: %s" % meta_path) if b"{{" in data: if PEDANTIC: raise RecipeError("found {{ in %s (Jinja templating not allowed)" % meta_path) else: data = render_jinja2(recipe_dir) else: data = data.decode("utf-8") check_dir_content(recipe_dir) for cfg in iter_cfgs(): meta = parse(data, cfg) validate_meta(meta) validate_files(recipe_dir, meta)