allow extended character set in annotations

INSTALL.md

@@ -40,7 +40,7 @@ Both methods are supported. However, for most users we _strongly_ recommend to b
   - Boost.Container: header-only
   - Boost.Test: header-only or (optionally) as a compiled library, *only used for unit testing*
   - Boost.Range: header-only, *only used for unit testing*
-- [BTAS](http://github.com/ValeevGroup/BTAS), tag db884b020b5c13c312c07df9d5c03cea2d65afb2 . If usable BTAS installation is not found, TiledArray will download and compile
+- [BTAS](http://github.com/ValeevGroup/BTAS), tag 242871710dabd5ef337e5253000d3e38c1d977ba . If usable BTAS installation is not found, TiledArray will download and compile
   BTAS from source. *This is the recommended way to compile BTAS for all users*.
 - [MADNESS](https://github.com/m-a-d-n-e-s-s/madness), tag 997e8b458c4234fb6c8c2781a5df59cb14b7e700 .
   Only the MADworld runtime and BLAS/LAPACK C API component of MADNESS is used by TiledArray.

external/versions.cmake

@@ -24,8 +24,8 @@ set(TA_TRACKED_MADNESS_PREVIOUS_TAG fae8081179b9d074968b08e064a32e3ca07ab0f1)
 set(TA_TRACKED_MADNESS_VERSION 0.10.1)
 set(TA_TRACKED_MADNESS_PREVIOUS_VERSION 0.10.1)
 
-set(TA_TRACKED_BTAS_TAG db884b020b5c13c312c07df9d5c03cea2d65afb2)
-set(TA_TRACKED_BTAS_PREVIOUS_TAG 1dc111d3efd6d6a4ba6f3b8cbc239315d0392121)
+set(TA_TRACKED_BTAS_TAG 242871710dabd5ef337e5253000d3e38c1d977ba)
+set(TA_TRACKED_BTAS_PREVIOUS_TAG db884b020b5c13c312c07df9d5c03cea2d65afb2)
 
 set(TA_TRACKED_CUTT_TAG 0e8685bf82910bc7435835f846e88f1b39f47f09)
 set(TA_TRACKED_CUTT_PREVIOUS_TAG 592198b93c93b7ca79e7900b9a9f2e79f9dafec3)

src/TiledArray/expressions/blk_tsr_expr.h

@@ -234,7 +234,7 @@ class BlkTsrExprBase : public Expr<Derived> {
   BlkTsrExprBase(reference array, const std::string& annotation,
                  const PairRange& bounds)
       : Expr_(), array_(array), annotation_(annotation) {
-    const auto rank = array.range().rank();
+    const auto rank = array.tiles_range().rank();
     lower_bound_.reserve(rank);
     upper_bound_.reserve(rank);
     int d = 0;

src/TiledArray/util/annotation.h

@@ -25,10 +25,14 @@
 
 #ifndef TILEDARRAY_ANNOTATION_H__INCLUDED
 #define TILEDARRAY_ANNOTATION_H__INCLUDED
+
+#include "TiledArray/error.h"
+
 #include <algorithm>
 #include <cstring>
+#include <sstream>
 #include <string>
-#include "TiledArray/error.h"
+#include <vector>
 
 namespace TiledArray::detail {
 
@@ -96,10 +100,12 @@ inline auto tokenize_index(const std::string& s, char delim) {
 /// TiledArray defines a string as being a valid index if each character is one
 /// of the following:
 ///
-/// - Roman letters A through Z (uppercase and lowercase are allowed)
-/// - Base 10 numbers 0 through 9
-/// - Whitespace
-/// - underscore (`_`), prime (<code>'</code>), comma (`,`), or semicolon (`;`)
+/// - Roman letters (`A..Z`, `a..z`)
+/// - decimal digits (`0..9`)
+/// - whitespace (` `)
+/// - comma (`,`)
+/// - semicolon (`;`)
+/// - any of the following characters: `'`_~!@#$%^&*-+./?:|<>[]{}()`
 ///
 /// Additionally the string can not:
 ///
@@ -124,7 +130,7 @@ inline bool is_valid_index(const std::string& idx) {
       "abcdefghijklmnopqrstuvwxyz"
       "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
       "1234567890"
-      " _',;";
+      ",; '`_~!@#$%^&*-+./?:|<>[]{}()";
   // Are valid characters
   for (const auto& c : idx)
     if (valid_chars.find(c) == std::string::npos) return false;

tests/annotation.cpp

@@ -1,6 +1,6 @@
-#include "unit_test_config.h"
 #include <TiledArray/util/annotation.h>
 #include <tiledarray.h>
+#include "unit_test_config.h"
 using namespace TiledArray::detail;
 
 namespace {
@@ -14,10 +14,10 @@ corr_map j_idx{{"j", "j"}, {" j", "j"}, {"j ", "j"}};
 corr_map k_idx{{"k", "k"}, {" k", "k"}, {"k ", "k"}};
 
 auto combine_maps(const corr_map& lhs, const corr_map& rhs,
-                  const std::string& joiner = ""){
+                  const std::string& joiner = "") {
   corr_map rv;
-  for(auto& [lidx, lcorr] : lhs){
-    for(auto& [ridx, rcorr] : rhs){
+  for (auto& [lidx, lcorr] : lhs) {
+    for (auto& [ridx, rcorr] : rhs) {
       rv[lidx + joiner + ridx] = lcorr + joiner + rcorr;
     }
   }
@@ -31,7 +31,7 @@ auto i_j_idx = combine_maps(i_idx, j_idx, ",");
 auto vov_idx = combine_maps(i_idx, j_idx, ";");
 auto vom_idx = combine_maps(k_idx, i_j_idx, ";");
 auto mov_idx = combine_maps(i_j_idx, k_idx, ";");
-} // namespace
+}  // namespace
 
 /* We need to remove whitespace from all types of indices: single character,
  * multiple character, multiple mode, and tensor-of-tensor. This test suite
@@ -48,42 +48,36 @@ auto mov_idx = combine_maps(i_j_idx, k_idx, ";");
  */
 BOOST_AUTO_TEST_SUITE(remove_whitespace_fxn)
 
-BOOST_AUTO_TEST_CASE(single_character){
-  for(auto& [idx, corr] : i_idx)
-    BOOST_CHECK(remove_whitespace(idx) == corr);
+BOOST_AUTO_TEST_CASE(single_character) {
+  for (auto& [idx, corr] : i_idx) BOOST_CHECK(remove_whitespace(idx) == corr);
 }
 
-BOOST_AUTO_TEST_CASE(multicharacter){
+BOOST_AUTO_TEST_CASE(multicharacter) {
   auto ij_idx = combine_maps(i_idx, j_idx);
 
-  for(auto& [idx, corr] : ij_idx)
-    BOOST_CHECK(remove_whitespace(idx) == corr);
+  for (auto& [idx, corr] : ij_idx) BOOST_CHECK(remove_whitespace(idx) == corr);
 }
 
-BOOST_AUTO_TEST_CASE(matrix_index){
-  for(auto& [idx, corr] : i_j_idx)
-    BOOST_CHECK(remove_whitespace(idx) == corr);
+BOOST_AUTO_TEST_CASE(matrix_index) {
+  for (auto& [idx, corr] : i_j_idx) BOOST_CHECK(remove_whitespace(idx) == corr);
 }
 
-BOOST_AUTO_TEST_CASE(tensor_index){
+BOOST_AUTO_TEST_CASE(tensor_index) {
   auto i_j_k_idx = combine_maps(i_j_idx, k_idx, ",");
-  for(auto& [idx, corr] : i_j_k_idx)
+  for (auto& [idx, corr] : i_j_k_idx)
     BOOST_CHECK(remove_whitespace(idx) == corr);
 }
 
-BOOST_AUTO_TEST_CASE(vector_of_vector_index){
-  for(auto& [idx, corr] : vov_idx)
-    BOOST_CHECK(remove_whitespace(idx) == corr);
+BOOST_AUTO_TEST_CASE(vector_of_vector_index) {
+  for (auto& [idx, corr] : vov_idx) BOOST_CHECK(remove_whitespace(idx) == corr);
 }
 
-BOOST_AUTO_TEST_CASE(vector_of_matrix){
-  for(auto& [idx, corr] : vom_idx)
-    BOOST_CHECK(remove_whitespace(idx) == corr);
+BOOST_AUTO_TEST_CASE(vector_of_matrix) {
+  for (auto& [idx, corr] : vom_idx) BOOST_CHECK(remove_whitespace(idx) == corr);
 }
 
-BOOST_AUTO_TEST_CASE(matrix_of_vector){
-  for(auto& [idx, corr] : mov_idx)
-    BOOST_CHECK(remove_whitespace(idx) == corr);
+BOOST_AUTO_TEST_CASE(matrix_of_vector) {
+  for (auto& [idx, corr] : mov_idx) BOOST_CHECK(remove_whitespace(idx) == corr);
 }
 
 BOOST_AUTO_TEST_SUITE_END()
@@ -95,7 +89,7 @@ BOOST_AUTO_TEST_SUITE_END()
  */
 BOOST_AUTO_TEST_SUITE(tokenize_index_fxn)
 
-BOOST_AUTO_TEST_CASE(split_on_comma){
+BOOST_AUTO_TEST_CASE(split_on_comma) {
   std::map<std::string, string_vector_t> input2corr{
       {"", string_vector_t{""}},
       {"hello world", string_vector_t{"hello world"}},
@@ -106,13 +100,12 @@ BOOST_AUTO_TEST_CASE(split_on_comma){
       {"hello,world ", string_vector_t{"hello", "world "}},
       {"hello world,", string_vector_t{"hello world", ""}},
       {"hello world, ", string_vector_t{"hello world", " "}},
-      {"1,2,3", string_vector_t{"1", "2", "3"}}
-  };
-  for(auto& [str, corr] : input2corr)
+      {"1,2,3", string_vector_t{"1", "2", "3"}}};
+  for (auto& [str, corr] : input2corr)
     BOOST_CHECK(tokenize_index(str, ',') == corr);
 }
 
-BOOST_AUTO_TEST_CASE(split_on_semicolon){
+BOOST_AUTO_TEST_CASE(split_on_semicolon) {
   std::map<std::string, string_vector_t> input2corr{
       {"", string_vector_t{""}},
       {"hello world", string_vector_t{"hello world"}},
@@ -123,9 +116,8 @@ BOOST_AUTO_TEST_CASE(split_on_semicolon){
       {"hello;world ", string_vector_t{"hello", "world "}},
       {"hello world;", string_vector_t{"hello world", ""}},
       {"hello world;", string_vector_t{"hello world", " "}},
-      {"1;2;3", string_vector_t{"1", "2", "3"}}
-  };
-  for(auto& [str, corr] : input2corr)
+      {"1;2;3", string_vector_t{"1", "2", "3"}}};
+  for (auto& [str, corr] : input2corr)
     BOOST_CHECK(tokenize_index(str, ';') == corr);
 }
 
@@ -139,29 +131,34 @@ BOOST_AUTO_TEST_SUITE_END()
  */
 BOOST_AUTO_TEST_SUITE(is_valid_index_fxn)
 
-BOOST_AUTO_TEST_CASE(valid_indices){
-  for(auto idx_set : {i_idx, i_j_idx, vov_idx, vom_idx, mov_idx}){
-    for(auto& [idx, corr] : idx_set) {
+BOOST_AUTO_TEST_CASE(valid_indices) {
+  for (auto idx_set : {i_idx, i_j_idx, vov_idx, vom_idx, mov_idx}) {
+    for (auto& [idx, corr] : idx_set) {
       BOOST_CHECK(is_valid_index(idx));
     }
   }
+  // all valid characters forming index name
+  BOOST_CHECK(
+      is_valid_index("abcdefghijklmnopqrstuvwxyz,ABCDEFGHIJKLMNOPQRSTUVWXYZ;'`_"
+                     "~!@#$%^&*-+.,/?:|<>[]{}"));
 }
 
-BOOST_AUTO_TEST_CASE(unallowed_character){
-  BOOST_CHECK(is_valid_index("i,&,j") == false);
+BOOST_AUTO_TEST_CASE(unallowed_character) {
+  BOOST_CHECK(is_valid_index("i,\",j") == false);
+  BOOST_CHECK(is_valid_index("i,\\,j") == false);
 }
 
-BOOST_AUTO_TEST_CASE(multiple_semicolons){
+BOOST_AUTO_TEST_CASE(multiple_semicolons) {
   BOOST_CHECK(is_valid_index("i;j;k") == false);
 }
 
-BOOST_AUTO_TEST_CASE(only_whitespace){
+BOOST_AUTO_TEST_CASE(only_whitespace) {
   BOOST_CHECK(is_valid_index("") == false);
   BOOST_CHECK(is_valid_index(" ") == false);
   BOOST_CHECK(is_valid_index("     ") == false);
 }
 
-BOOST_AUTO_TEST_CASE(empty_index_name){
+BOOST_AUTO_TEST_CASE(empty_index_name) {
   BOOST_CHECK(is_valid_index("i,") == false);
   BOOST_CHECK(is_valid_index(",i") == false);
   BOOST_CHECK(is_valid_index("i,,j") == false);
@@ -182,19 +179,17 @@ BOOST_AUTO_TEST_SUITE_END()
  */
 BOOST_AUTO_TEST_SUITE(is_tot_index_fxn)
 
-BOOST_AUTO_TEST_CASE(valid_but_not_tot){
-  for(auto x : {i_idx, i_j_idx})
-    for(auto idx : x)
-      BOOST_CHECK(is_tot_index(idx.first) == false);
+BOOST_AUTO_TEST_CASE(valid_but_not_tot) {
+  for (auto x : {i_idx, i_j_idx})
+    for (auto idx : x) BOOST_CHECK(is_tot_index(idx.first) == false);
 }
 
-BOOST_AUTO_TEST_CASE(valid_tot_idx){
-  for(auto x : {vov_idx, vom_idx, mov_idx})
-    for(auto idx : x)
-      BOOST_CHECK(is_tot_index(idx.first));
+BOOST_AUTO_TEST_CASE(valid_tot_idx) {
+  for (auto x : {vov_idx, vom_idx, mov_idx})
+    for (auto idx : x) BOOST_CHECK(is_tot_index(idx.first));
 }
 
-BOOST_AUTO_TEST_CASE(not_valid_idx){
+BOOST_AUTO_TEST_CASE(not_valid_idx) {
   BOOST_CHECK(is_tot_index("") == false);
   BOOST_CHECK(is_tot_index(";") == false);
 }
@@ -208,32 +203,29 @@ BOOST_AUTO_TEST_SUITE_END()
  */
 BOOST_AUTO_TEST_SUITE(split_index_fxn)
 
-BOOST_AUTO_TEST_CASE(invalid_idx){
-  if(TiledArray::get_default_world().nproc() == 1)
+BOOST_AUTO_TEST_CASE(invalid_idx) {
+  if (TiledArray::get_default_world().nproc() == 1)
     BOOST_CHECK_THROW(split_index("i,"), TiledArray::Exception);
 }
 
-BOOST_AUTO_TEST_CASE(non_tot){
+BOOST_AUTO_TEST_CASE(non_tot) {
   std::map<std::string, std::pair<string_vector_t, string_vector_t>> inputs{
       {"i", {string_vector_t{"i"}, string_vector_t{}}},
-      {"i,j", {string_vector_t{"i","j"}, string_vector_t{}}},
-      {"i,j,k", {string_vector_t{"i", "j", "k"}, string_vector_t{}}}
-  };
+      {"i,j", {string_vector_t{"i", "j"}, string_vector_t{}}},
+      {"i,j,k", {string_vector_t{"i", "j", "k"}, string_vector_t{}}}};
 
-  for(auto& [idx, corr] : inputs) {
+  for (auto& [idx, corr] : inputs) {
     BOOST_CHECK(split_index(idx) == corr);
   }
 }
 
-BOOST_AUTO_TEST_CASE(tot){
+BOOST_AUTO_TEST_CASE(tot) {
   std::map<std::string, std::pair<string_vector_t, string_vector_t>> inputs{
       {"i;j", {string_vector_t{"i"}, string_vector_t{"j"}}},
-      {"i,j;k", {string_vector_t{"i","j"}, string_vector_t{"k"}}},
-      {"i;j,k", {string_vector_t{"i"}, string_vector_t{"j", "k"}}}
-  };
+      {"i,j;k", {string_vector_t{"i", "j"}, string_vector_t{"k"}}},
+      {"i;j,k", {string_vector_t{"i"}, string_vector_t{"j", "k"}}}};
 
-  for(auto& [idx, corr] : inputs)
-    BOOST_CHECK(split_index(idx) == corr);
+  for (auto& [idx, corr] : inputs) BOOST_CHECK(split_index(idx) == corr);
 }
 
 BOOST_AUTO_TEST_SUITE_END()