Fix Bibtex rendering in reports

docs/outputs.rst

@@ -546,7 +546,7 @@ An example report
   A two-stage masking procedure was applied, in which a liberal mask (including voxels with good data in at least the first echo) was used for optimal combination, T2*/S0 estimation, and denoising, while a more conservative mask (restricted to voxels with good data in at least the first three echoes) was used for the component classification procedure.
   Multi-echo data were then optimally combined using the T2* combination method \\citep{posse1999enhancement}.
   Next, components were manually classified as BOLD (TE-dependent), non-BOLD (TE-independent), or uncertain (low-variance).
-  This workflow used numpy \\citep{van2011numpy}, scipy \\citep{virtanen2020scipy}, pandas \\citep{mckinney2010data,reback2020pandas}, scikit-learn \\citep{pedregosa2011scikit}, nilearn, bokeh \\citep{bokehmanual}, matplotlib \\citep{Hunter:2007}, and nibabel \\citep{brett_matthew_2019_3233118}.
+  This workflow used numpy \\citep{van2011numpy}, scipy \\citep{virtanen2020scipy}, pandas \\citep{mckinney2010data,reback2020pandas}, scikit-learn \\citep{pedregosa2011scikit}, nilearn, bokeh \\citep{bokehmanual}, matplotlib \\citep{Hunter2007}, and nibabel \\citep{brett_matthew_2019_3233118}.
   This workflow also used the Dice similarity index \\citep{dice1945measures,sorensen1948method}.
 
   References

pyproject.toml

@@ -28,6 +28,8 @@ dependencies = [
     "nilearn>=0.7",
     "numpy>=1.16",
     "pandas>=2.0",
+    "pybtex",
+    "pybtex-apa-style",
     "scikit-learn>=0.21",
     "scipy>=1.2.0",
     "threadpoolctl",
@@ -48,6 +50,7 @@ doc = [
     "sphinx-argparse",
     "sphinxcontrib-bibtex",
 ]
+
 tests = [
     "codecov",
     "coverage",

tedana/bibtex.py

@@ -123,9 +123,9 @@ def find_citations(description):
     all_citations : :obj:`list` of :obj:`str`
         A list of all identifiers for citations.
     """
-    paren_citations = re.findall(r"\\citep{([a-zA-Z0-9,/\.]+)}", description)
-    intext_citations = re.findall(r"\\cite{([a-zA-Z0-9,/\.]+)}", description)
-    inparen_citations = re.findall(r"\\citealt{([a-zA-Z0-9,/\.]+)}", description)
+    paren_citations = re.findall(r"\\citep{([a-zA-Z0-9,_/\.]+)}", description)
+    intext_citations = re.findall(r"\\cite{([a-zA-Z0-9,_/\.]+)}", description)
+    inparen_citations = re.findall(r"\\citealt{([a-zA-Z0-9,_/\.]+)}", description)
     all_citations = ",".join(paren_citations + intext_citations + inparen_citations)
     all_citations = all_citations.split(",")
     all_citations = sorted(list(set(all_citations)))

tedana/reporting/data/html/report_body_template.html

@@ -24,6 +24,10 @@
     width: 80%;
   }
 
+  .references ul {
+    line-height: 150%;
+  }
+
   .carpet-wrapper {
     margin-top: 30px;
   }
@@ -177,9 +181,12 @@ <h1>Info</h1>
   <div class="about">
     <h1>About tedana</h1>
     $about
-
+  </div>
+  <div class="content references">
     <h1>References</h1>
-    $references
+    <ul>
+      $references
+    </ul>
   </div>
 </div>
 

tedana/reporting/html_report.py

@@ -1,13 +1,16 @@
 """Build HTML reports for tedana."""
 import logging
 import os
+import re
 from os.path import join as opj
 from pathlib import Path
 from string import Template
 
 import pandas as pd
 from bokeh import __version__ as bokehversion
 from bokeh import embed, layouts, models
+from pybtex.database.input import bibtex
+from pybtex.plugin import find_plugin
 
 from tedana import __version__
 from tedana.io import load_json
@@ -16,6 +19,60 @@
 LGR = logging.getLogger("GENERAL")
 
 
+APA = find_plugin("pybtex.style.formatting", "apa")()
+HTML = find_plugin("pybtex.backends", "html")()
+
+
+def _bib2html(bibliography):
+    parser = bibtex.Parser()
+    bibliography = parser.parse_file(bibliography)
+    formatted_bib = APA.format_bibliography(bibliography)
+    bibliography_str = "".join(f"<li>{entry.text.render(HTML)}</li>" for entry in formatted_bib)
+    return bibliography_str, bibliography
+
+
+def _cite2html(bibliography, citekey):
+    # Make a list of citekeys and separete double citations
+    citekey_list = citekey.split(",") if "," in citekey else [citekey]
+
+    for idx, key in enumerate(citekey_list):
+        # Get first author
+        first_author = bibliography.entries[key].persons["author"][0]
+
+        # Keep surname only (whatever is before the comma, if there is a comma)
+        if "," in str(first_author):
+            first_author = str(first_author).split(",")[0]
+
+        # Get publication year
+        pub_year = bibliography.entries[key].fields["year"]
+
+        # Return complete citation
+        if idx == 0:
+            citation = f"{first_author} et al. {pub_year}"
+        else:
+            citation += f", {first_author} et al. {pub_year}"
+
+    return citation
+
+
+def _inline_citations(text, bibliography):
+    # Find all \citep
+    matches = re.finditer(r"\\citep{(.*?)}", text)
+    citations = [(match.start(), match.group(1)) for match in matches]
+
+    updated_text = text
+
+    for citation in citations:
+        citekey = citation[1]
+        matched_string = "\\citep{" + citekey + "}"
+
+        # Convert citation form latex to html
+        html_citation = f"({_cite2html(bibliography, citekey)})"
+        updated_text = updated_text.replace(matched_string, html_citation, 1)
+
+    return updated_text
+
+
 def _generate_buttons(out_dir, io_generator):
     resource_path = Path(__file__).resolve().parent.joinpath("data", "html")
 
@@ -85,6 +142,12 @@ def _update_template_bokeh(bokeh_id, info_table, about, prefix, references, boke
     # Initial carpet plot (default one)
     initial_carpet = f"./figures/{prefix}carpet_optcom.svg"
 
+    # Convert bibtex to html
+    references, bibliography = _bib2html(references)
+
+    # Update inline citations
+    about = _inline_citations(about, bibliography)
+
     body_template_name = "report_body_template.html"
     body_template_path = resource_path.joinpath(body_template_name)
     with open(str(body_template_path)) as body_file:
@@ -273,8 +336,7 @@ def get_elbow_val(elbow_prefix):
     with open(opj(io_generator.out_dir, f"{io_generator.prefix}report.txt"), "r+") as f:
         about = f.read()
 
-    with open(opj(io_generator.out_dir, f"{io_generator.prefix}references.bib")) as f:
-        references = f.read()
+    references = opj(io_generator.out_dir, f"{io_generator.prefix}references.bib")
 
     # Read info table
     data_descr_path = io_generator.get_name("data description json")