correctionlib 2.4.0__tar.gz → 2.5.0__tar.gz

{correctionlib-2.4.0 → correctionlib-2.5.0}/.github/workflows/ci.yml

@@ -30,7 +30,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: ["3.7", "3.11"]
+        python-version: ["3.7", "3.11", "3.12"]
         runs-on: [ubuntu-latest, macos-latest, windows-latest]
 
         include:

{correctionlib-2.4.0 → correctionlib-2.5.0}/.github/workflows/wheels.yml

@@ -42,7 +42,7 @@ jobs:
         fetch-tags: true
         fetch-depth: 0
 
-    - uses: pypa/cibuildwheel@v2.14.1
+    - uses: pypa/cibuildwheel@v2.16.5
       env:
         MACOSX_DEPLOYMENT_TARGET: 10.14
 
@@ -62,7 +62,7 @@ jobs:
         name: artifact
         path: dist
 
-    - uses: pypa/gh-action-pypi-publish@v1.8.10
+    - uses: pypa/gh-action-pypi-publish@v1.8.11
       with:
         user: __token__
         password: ${{ secrets.pypi_token }}

{correctionlib-2.4.0 → correctionlib-2.5.0}/.pre-commit-config.yaml

@@ -41,4 +41,4 @@ repos:
   hooks:
   - id: mypy
     files: src
-    additional_dependencies: [uhi, pydantic<2, numpy, types-setuptools]
+    additional_dependencies: [uhi, pydantic, numpy, types-setuptools]

{correctionlib-2.4.0/src/correctionlib.egg-info → correctionlib-2.5.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: correctionlib
-Version: 2.4.0
+Version: 2.5.0
 Summary: A generic correction library
 Home-page: https://github.com/cms-nanoAOD/correctionlib
 Author: Nick Smith
@@ -17,14 +17,13 @@ Classifier: License :: OSI Approved :: BSD License
 Classifier: Programming Language :: Python
 Classifier: Programming Language :: Python :: 3
 Classifier: Development Status :: 5 - Production/Stable
-Requires-Python: >=3.6
+Requires-Python: >=3.7
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: numpy>=1.13.3
-Requires-Dist: typing; python_version < "3.5"
 Requires-Dist: typing-extensions; python_version < "3.8"
 Requires-Dist: dataclasses; python_version < "3.7"
-Requires-Dist: pydantic<2,>=1.7.3
+Requires-Dist: pydantic>=2
 Requires-Dist: rich
 Requires-Dist: packaging
 Provides-Extra: test
@@ -32,7 +31,7 @@ Requires-Dist: pytest>=4.6; extra == "test"
 Requires-Dist: scipy; extra == "test"
 Requires-Dist: awkward>=2.2.2; python_version > "3.7" and extra == "test"
 Requires-Dist: awkward<2; python_version <= "3.7" and extra == "test"
-Requires-Dist: dask-awkward; python_version > "3.7" and extra == "test"
+Requires-Dist: dask-awkward>=2024.1.1; python_version > "3.7" and extra == "test"
 Provides-Extra: dev
 Requires-Dist: pytest>=4.6; extra == "dev"
 Requires-Dist: pre-commit; extra == "dev"

{correctionlib-2.4.0 → correctionlib-2.5.0}/data/conversion.py

@@ -33,7 +33,7 @@ def build_formula(sf):
 
     value = sf.iloc[0]["formula"]
     if "x" in value:
-        return Formula.parse_obj(
+        return Formula.model_validate(
             {
                 "nodetype": "formula",
                 "expression": value,
@@ -49,7 +49,7 @@ def build_formula(sf):
 
 def build_discrbinning(sf):
     edges = sorted(set(sf["discrMin"]) | set(sf["discrMax"]))
-    return Binning.parse_obj(
+    return Binning.model_validate(
         {
             "nodetype": "binning",
             "input": "discriminant",
@@ -65,7 +65,7 @@ def build_discrbinning(sf):
 
 def build_ptbinning(sf):
     edges = sorted(set(sf["ptMin"]) | set(sf["ptMax"]))
-    return Binning.parse_obj(
+    return Binning.model_validate(
         {
             "nodetype": "binning",
             "input": "pt",
@@ -81,7 +81,7 @@ def build_ptbinning(sf):
 
 def build_etabinning(sf):
     edges = sorted(set(sf["etaMin"]) | set(sf["etaMax"]))
-    return Binning.parse_obj(
+    return Binning.model_validate(
         {
             "nodetype": "binning",
             "input": "abseta",
@@ -97,7 +97,7 @@ def build_etabinning(sf):
 
 def build_flavor(sf):
     keys = sorted(sf["jetFlavor"].unique())
-    return Category.parse_obj(
+    return Category.model_validate(
         {
             "nodetype": "category",
             "input": "flavor",
@@ -111,7 +111,7 @@ def build_flavor(sf):
 
 def build_systs(sf):
     keys = list(sf["sysType"].unique())
-    return Category.parse_obj(
+    return Category.model_validate(
         {
             "nodetype": "category",
             "input": "systematic",
@@ -123,7 +123,7 @@ def build_systs(sf):
     )
 
 
-corr2 = Correction.parse_obj(
+corr2 = Correction.model_validate(
     {
         "version": 1,
         "name": "DeepCSV_2016LegacySF",
@@ -149,11 +149,11 @@ corr2 = Correction.parse_obj(
 )
 
 
-sf = requests.get(f"{examples}/EIDISO_WH_out.histo.json").json()
+sf = requests.get(f"{examples}/EIDISO_WH_out.histo.json").model_dump_json()
 
 
 def build_syst(sf):
-    return Category.parse_obj(
+    return Category.model_validate(
         {
             "nodetype": "category",
             "input": "systematic",
@@ -187,7 +187,7 @@ def build_pts(sf):
         edges.append(hi)
         content.append(build_syst(data))
 
-    return Binning.parse_obj(
+    return Binning.model_validate(
         {
             "nodetype": "binning",
             "input": "pt",
@@ -212,7 +212,7 @@ def build_etas(sf):
         if not found:
             raise ValueError("eta edges not in binning?")
 
-    return Binning.parse_obj(
+    return Binning.model_validate(
         {
             "nodetype": "binning",
             "input": "eta",
@@ -223,7 +223,7 @@ def build_etas(sf):
     )
 
 
-corr3 = Correction.parse_obj(
+corr3 = Correction.model_validate(
     {
         "version": 1,
         "name": "EIDISO_WH_out",
@@ -239,7 +239,7 @@ corr3 = Correction.parse_obj(
 )
 
 
-cset = CorrectionSet.parse_obj(
+cset = CorrectionSet.model_validate(
     {
         "schema_version": VERSION,
         "corrections": [
@@ -251,4 +251,4 @@ cset = CorrectionSet.parse_obj(
 )
 
 with gzip.open("data/examples.json.gz", "wt") as fout:
-    fout.write(cset.json(exclude_unset=True))
+    fout.write(cset.model_dump_json(exclude_unset=True))

{correctionlib-2.4.0 → correctionlib-2.5.0}/include/correction.h

@@ -188,7 +188,7 @@ struct _UniformBins {
 class Binning {
   public:
     Binning(const JSONObject& json, const Correction& context);
-    const Content& child(const std::vector<Variable::Type>& values) const;
+    double evaluate(const std::vector<Variable::Type>& values) const;
 
   private:
     std::variant<_UniformBins, _NonUniformBins> bins_; // bin edges
@@ -209,7 +209,7 @@ class MultiBinning {
   public:
     MultiBinning(const JSONObject& json, const Correction& context);
     size_t ndimensions() const { return axes_.size(); };
-    const Content& child(const std::vector<Variable::Type>& values) const;
+    double evaluate(const std::vector<Variable::Type>& values) const;
 
   private:
     size_t nbins(size_t dimension) const;
@@ -222,7 +222,7 @@ class MultiBinning {
 class Category {
   public:
     Category(const JSONObject& json, const Correction& context);
-    const Content& child(const std::vector<Variable::Type>& values) const;
+    double evaluate(const std::vector<Variable::Type>& values) const;
 
   private:
     typedef std::map<int, Content> IntMap;

{correctionlib-2.4.0 → correctionlib-2.5.0}/pyproject.toml

@@ -15,8 +15,8 @@ write_to = "src/correctionlib/version.py"
 skip = ["pp*-*"]
 test-extras = "test"
 test-command = "python -m pytest {package}/tests"
-# update skip when numpy wheels become available
-test-skip = ["pp*-*", "*-musllinux_*", "cp310-win32", "cp311-win32", "cp38-manylinux_i686", "cp39-manylinux_i686", "cp310-manylinux_i686", "cp311-manylinux_i686"]
+# When cpython 3.7 is dropped we can ignore i686 (numpy not built)
+test-skip = ["pp*-*", "*-musllinux_*", "cp3{10,11,12}-win32", "cp3{8,9,10,11,12}-manylinux_i686"]
 manylinux-x86_64-image = "manylinux2014"
 manylinux-i686-image = "manylinux2014"
 

{correctionlib-2.4.0 → correctionlib-2.5.0}/setup.cfg

@@ -22,13 +22,12 @@ classifiers =
 	Development Status :: 5 - Production/Stable
 
 [options]
-python_requires = >=3.6
+python_requires = >=3.7
 install_requires = 
 	numpy >=1.13.3
-	typing; python_version<"3.5"
 	typing-extensions;python_version<"3.8"
 	dataclasses;python_version<"3.7"
-	pydantic <2,>=1.7.3
+	pydantic >=2
 	rich
 	packaging
 
@@ -38,7 +37,7 @@ test =
 	scipy
 	awkward >=2.2.2;python_version>"3.7"
 	awkward <2;python_version<="3.7"
-	dask-awkward;python_version>"3.7"
+	dask-awkward >=2024.1.1;python_version>"3.7"
 dev = 
 	pytest >=4.6
 	pre-commit

{correctionlib-2.4.0 → correctionlib-2.5.0}/src/correction.cc

@@ -103,7 +103,7 @@ std::string_view JSONObject::getRequired<std::string_view>(const char * key) con
 
 namespace {
   Content resolve_content(const rapidjson::Value& json, const Correction& context) {
-    if ( json.IsDouble() ) { return json.GetDouble(); }
+    if ( json.IsNumber() ) { return json.GetDouble(); }
     else if ( json.IsObject() && json.HasMember("nodetype") ) {
       auto obj = JSONObject(json.GetObject());
       auto type = obj.getRequired<std::string_view>("nodetype");
@@ -120,28 +120,12 @@ namespace {
   }
 
   struct node_evaluate {
-    double operator() (double node) { return node; };
-    double operator() (const Binning& node) {
-      return std::visit(*this, node.child(values));
-    };
-    double operator() (const MultiBinning& node) {
-      return std::visit(*this, node.child(values));
-    };
-    double operator() (const Category& node) {
-      return std::visit(*this, node.child(values));
-    };
-    double operator() (const Formula& node) {
-      return node.evaluate(values);
-    };
-    double operator() (const FormulaRef& node) {
-      return node.evaluate(values);
-    };
-    double operator() (const Transform& node) {
-      return node.evaluate(values);
-    };
-    double operator() (const HashPRNG& node) {
+    double operator() (double node) { return node; }
+
+    template <class Node>
+    double operator() (const Node &node) {
       return node.evaluate(values);
-    };
+    }
 
     const std::vector<Variable::Type>& values;
   };
@@ -468,10 +452,12 @@ Binning::Binning(const JSONObject& json, const Correction& context)
   contents_.push_back(std::move(default_value));
 }
 
-const Content& Binning::child(const std::vector<Variable::Type>& values) const {
+double Binning::evaluate(const std::vector<Variable::Type>& values) const
+{
   double value = std::get<double>(values[variableIdx_]);
   std::size_t binIdx = find_bin_idx(value, bins_, flow_, variableIdx_, "Binning");
-  return contents_[binIdx];
+  const Content& child = contents_[binIdx];
+  return std::visit(node_evaluate{values}, child);
 }
 
 MultiBinning::MultiBinning(const JSONObject& json, const Correction& context)
@@ -540,7 +526,8 @@ MultiBinning::MultiBinning(const JSONObject& json, const Correction& context)
   }
 }
 
-const Content& MultiBinning::child(const std::vector<Variable::Type>& values) const {
+double MultiBinning::evaluate(const std::vector<Variable::Type>& values) const
+{
   size_t idx {0};
   size_t localidx {0};
   size_t dim {0};
@@ -549,12 +536,13 @@ const Content& MultiBinning::child(const std::vector<Variable::Type>& values) co
     double value = std::get<double>(values[variableIdx]);
     localidx = find_bin_idx(value, edgesVariant, flow_, variableIdx, "MultiBinning");
     if ( localidx == nbins(dim) ) // find_bin_idx is indicating we need to return the default value
-      return content_.back();
+      return std::visit(node_evaluate{values}, content_.back());
     idx += localidx * stride;
     ++dim;
   }
 
-  return content_.at(idx);
+  const Content& child = content_.at(idx);
+  return std::visit(node_evaluate{values}, child);
 }
 
 size_t MultiBinning::nbins(size_t dimension) const
@@ -581,13 +569,13 @@ Category::Category(const JSONObject& json, const Correction& context)
     }
     if ( kv_pair["key"].IsString() ) {
       if ( variable.type() != Variable::VarType::string ) {
-        throw std::runtime_error("Category got a key not of type string, but its input is string type");
+        throw std::runtime_error("Category got a key of type string, but its input is type " + variable.typeStr());
       }
       std::get<StrMap>(map_).try_emplace(kv_pair["key"].GetString(), resolve_content(kv_pair["value"], context));
     }
     else if ( kv_pair["key"].IsInt() ) {
       if ( variable.type() != Variable::VarType::integer ) {
-        throw std::runtime_error("Category got a key not of type int, but its input is int type");
+        throw std::runtime_error("Category got a key of type int, but its input is type " + variable.typeStr());
       }
       std::get<IntMap>(map_).try_emplace(kv_pair["key"].GetInt(), resolve_content(kv_pair["value"], context));
     }
@@ -602,13 +590,14 @@ Category::Category(const JSONObject& json, const Correction& context)
   }
 }
 
-const Content& Category::child(const std::vector<Variable::Type>& values) const {
+double Category::evaluate(const std::vector<Variable::Type>& values) const {
+  const Content* child = nullptr;
   if ( auto pval = std::get_if<std::string>(&values[variableIdx_]) ) {
     try {
-      return std::get<StrMap>(map_).at(*pval);
+      child = &std::get<StrMap>(map_).at(*pval);
     } catch (std::out_of_range& ex) {
       if ( default_ ) {
-        return *default_;
+        child = default_.get();
       }
       else {
         throw std::out_of_range("Index not available in Category for input argument " + std::to_string(variableIdx_) + " val: " + *pval);
@@ -617,17 +606,20 @@ const Content& Category::child(const std::vector<Variable::Type>& values) const
   }
   else if ( auto pval = std::get_if<int>(&values[variableIdx_]) ) {
     try {
-      return std::get<IntMap>(map_).at(*pval);
+      child = &std::get<IntMap>(map_).at(*pval);
     } catch (std::out_of_range& ex) {
       if ( default_ ) {
-        return *default_;
+        child = default_.get();
       }
       else {
         throw std::out_of_range("Index not available in Category for input argument " + std::to_string(variableIdx_) + " val: " + std::to_string(*pval));
       }
     }
+  } else {
+    throw std::runtime_error("Invalid variable type");
   }
-  throw std::runtime_error("Invalid variable type");
+
+  return std::visit(node_evaluate{values}, *child);
 }
 
 Correction::Correction(const JSONObject& json) :

{correctionlib-2.4.0 → correctionlib-2.5.0}/src/correctionlib/JSONEncoder.py

@@ -29,9 +29,8 @@ def write(data: Any, fname: str, **kwargs: Any) -> None:
 def dumps(data: Any, sort_keys: bool = False, **kwargs: Any) -> str:
     """Help function to quickly dump dictionary formatted by JSONEncoder."""
     if isinstance(data, pydantic.BaseModel):  # for pydantic
-        return data.json(cls=JSONEncoder, exclude_unset=True, **kwargs)
-    else:  # for standard data structures
-        return json.dumps(data, cls=JSONEncoder, sort_keys=sort_keys, **kwargs)
+        data = data.model_dump(mode="json", exclude_unset=True)
+    return json.dumps(data, cls=JSONEncoder, sort_keys=sort_keys, **kwargs)
 
 
 class JSONEncoder(json.JSONEncoder):

{correctionlib-2.4.0 → correctionlib-2.5.0}/src/correctionlib/__init__.py

@@ -2,15 +2,10 @@ import sys
 
 if sys.platform.startswith("win32"):
     import ctypes
-    import os.path
 
-    import pkg_resources
+    from .util import this_module_path
 
-    ctypes.CDLL(
-        pkg_resources.resource_filename(
-            "correctionlib", os.path.join("lib", "correctionlib.dll")
-        )
-    )
+    ctypes.CDLL(str(this_module_path() / "lib" / "correctionlib.dll"))
 
 
 from .binding import register_pyroot_binding

correctionlib-2.5.0/src/correctionlib/binding.py

@@ -0,0 +1,20 @@
+def register_pyroot_binding() -> None:
+    import sys
+
+    from cppyy import gbl
+
+    from .util import this_module_path
+
+    base_path = this_module_path()
+    lib = base_path / "lib"
+
+    # maybe not the most robust solution?
+    if sys.platform.startswith("win32"):
+        lib = lib / "correctionlib.dll"
+    elif sys.platform.startswith("darwin"):
+        lib = lib / "libcorrectionlib.dylib"
+    else:
+        lib = lib / "libcorrectionlib.so"
+    gbl.gSystem.Load(lib)
+    gbl.gInterpreter.AddIncludePath(base_path / "include")
+    gbl.gROOT.ProcessLine('#include "correction.h"')

{correctionlib-2.4.0 → correctionlib-2.5.0}/src/correctionlib/cli.py

@@ -101,9 +101,9 @@ def merge(console: Console, args: argparse.Namespace) -> int:
             cset.compound_corrections.append(corr2)
     cset.description = "Merged from " + " ".join(args.files)
     if args.format == "compact":
-        sys.stdout.write(cset.json())
+        sys.stdout.write(cset.model_dump_json())
     elif args.format == "indented":
-        sys.stdout.write(cset.json(indent=4) + "\n")
+        sys.stdout.write(cset.model_dump_json(indent=4) + "\n")
     elif args.format == "pretty":
         from correctionlib.JSONEncoder import dumps
 
@@ -130,27 +130,26 @@ def setup_merge(subparsers: Any) -> None:
 
 
 def config(console: Console, args: argparse.Namespace) -> int:
-    import pkg_resources
+    from .util import this_module_path
 
-    incdir = pkg_resources.resource_filename("correctionlib", "include")
-    libdir = pkg_resources.resource_filename("correctionlib", "lib")
+    base_dir = this_module_path()
+    incdir = base_dir / "include"
+    libdir = base_dir / "lib"
     out = []
     if args.version:
         out.append(correctionlib.version.version)
     if args.incdir:
-        out.append(incdir)
+        out.append(str(incdir))
     if args.cflags:
         out.append(f"-std=c++17 -I{incdir}")
     if args.libdir:
-        out.append(libdir)
+        out.append(str(libdir))
     if args.ldflags:
         out.append(f"-L{libdir} -lcorrectionlib")
     if args.rpath:
         out.append(f"-Wl,-rpath,{libdir}")
     if args.cmake:
-        out.append(
-            f"-Dcorrectionlib_DIR={pkg_resources.resource_filename('correctionlib', 'cmake')}"
-        )
+        out.append(f"-Dcorrectionlib_DIR={base_dir / 'cmake'}")
     console.out(" ".join(out), highlight=False)
     return 0
 

{correctionlib-2.4.0 → correctionlib-2.5.0}/src/correctionlib/convert.py

@@ -82,7 +82,7 @@ def from_histogram(
         axname = getattr(
             axis, "name", f"axis{pos}" if axis_names is None else axis_names[pos]
         )
-        return Variable.parse_obj(
+        return Variable.model_validate(
             {
                 "type": axtype,
                 "name": axname,
@@ -120,7 +120,7 @@ def from_histogram(
     ) -> Content:
         vartype = variables[0].type
         if vartype in {"string", "int"}:
-            return Category.parse_obj(
+            return Category.model_validate(
                 {
                     "nodetype": "category",
                     "input": variables[0].name,
@@ -142,7 +142,7 @@ def from_histogram(
                 break
             i += 1
         if i > 1:
-            return MultiBinning.parse_obj(
+            return MultiBinning.model_validate(
                 {
                     "nodetype": "multibinning",
                     "edges": [edges(ax) for ax in axes[:i]],
@@ -156,7 +156,7 @@ def from_histogram(
                     "flow": flow,
                 }
             )
-        return Binning.parse_obj(
+        return Binning.model_validate(
             {
                 "nodetype": "binning",
                 "input": variables[0].name,
@@ -171,7 +171,7 @@ def from_histogram(
             }
         )
 
-    return Correction.parse_obj(
+    return Correction.model_validate(
         {
             "version": 0,
             "name": getattr(hist, "name", "unknown"),