finding misextracted subscripts in pdfs, using PdfTextStream

subscripts.java

@Test
    public void testSubscripts() throws Exception {

        final Pattern SUBSCRIPTS = Pattern.compile("^[ \\d]{10,1000}$");

        File ROOT = new File(
                        "/Volumes/scratch/richarde/pdfs/201307/");
        for (File pdf : ROOT.listFiles()) {
            if (pdf.getName().endsWith(".pdf")) {
                try {

                    // System.out.println("reading " + pdf.getName());
                    PDFTextStream pdfts = new PDFTextStream(pdf);
                    StringBuilder text = new StringBuilder(1024);
                    pdfts.pipe(new OutputTarget(text));
                    pdfts.close();
                    String previous = "";
                    for (String line : text.toString().split("\n")) {

                        Matcher m = SUBSCRIPTS.matcher(line);
                        while (m.find()) {
                            System.out
                                    .println(pdf.getName()
                                            + "--------------------------------------\n"
                                            + previous + "\n" + line + "\n");
                        }

                        previous = line;
                    }
                } catch (Exception e) {
                    System.err.println(e);
                }
            }
        }
    }

subscripts.txt

http://www.ncbi.nlm.nih.gov/pubmed/?term=10720617
fluid 119.0 mM NaCl, 3.3 mM KCl, 1.3 mM CaCl  , 1.2
                                                  2

http://www.ncbi.nlm.nih.gov/pubmed/?term=10720617
mM MgCl  , 0.5 mM Na  HPO , 21.0 mM NaHCO , 3.4
2                                    24                                                     3

http://www.ncbi.nlm.nih.gov/pubmed/?term=10720617
blockade experiments was carried out using CoCl   Co2q.
                                                  2

http://www.ncbi.nlm.nih.gov/pubmed/?term=11166682
by  briefly  immersing  the  sections  in  0.05%  OsO .  Pro-
                                                  4


http://www.ncbi.nlm.nih.gov/pubmed/?term=9570713
mM  KH PO ,  136  mM  NaCl,  8  mM  Na HPO ).  Slides  were
24                                                                                   24


http://www.ncbi.nlm.nih.gov/pubmed/?term=9570713
aline  activates  vasopressin  neurons  via a -receptor-mediated  Ca21
                                                  1



http://www.ncbi.nlm.nih.gov/pubmed/?term=10708688
pCO   hypercapnia. which is metabolically compensated
               2

http://www.ncbi.nlm.nih.gov/pubmed/?term=10708688
artery ligation did not affect the arterial  pCO   w10x. The
                                                  2

http://www.ncbi.nlm.nih.gov/pubmed/?term=12900283
electron-donating group, p-Me N-, -OMe or -OH, which
                                                  2

http://www.ncbi.nlm.nih.gov/pubmed/?term=12900283
residual  protons  in  CDCl    unless  otherwise  mentioned.
                                                  3

http://www.ncbi.nlm.nih.gov/pubmed/?term=12900283
NaCNBH  (300 mg, 4.8 mmol). The resulting mixture was
                               3

http://www.ncbi.nlm.nih.gov/pubmed/?term=12900283
o)fluorene (94%). 1H NMR (200 MHz, CDCl ):    3.09 (s,
                                                  3

http://www.ncbi.nlm.nih.gov/pubmed/?term=12900283
MHz,  CDCl ):      45.43,  114.04,  119.74,  120.60,  120.98,
                                         3

http://www.ncbi.nlm.nih.gov/pubmed/?term=12900283
ofluorene. 1H NMR (200 MHz, CDCl ):    3.05 (s, 6H), 3.89
                                                  3

http://www.ncbi.nlm.nih.gov/pubmed/?term=12900283
ofluorene. 1H NMR (200 MHz, CDCl ):    3.01 (s, 6H), 4.01
                                                  3

http://www.ncbi.nlm.nih.gov/pubmed/?term=23015449
taining the following (in mM): 139 NaCl, 12 D-glucose, 17 NaHCO , 3
                                                  3

http://www.ncbi.nlm.nih.gov/pubmed/?term=23015449
mPSCs contained the following (in mM): 139 NaCl, 5 KCl, 17 NaHCO ,
                                                  3

http://www.ncbi.nlm.nih.gov/pubmed/?term=23015449
Na ATP, and 0.1 MgGTP; pH was adjusted to 7.3 with KOH. The junc-
         2

http://www.ncbi.nlm.nih.gov/pubmed/?term=23015449
when motoneurons were held at    70 mV and with 1 mM MgCl  in the
                                                  2

http://www.ncbi.nlm.nih.gov/pubmed/?term=218204
Na+ Na+  K+         Na+       Na+  K+          Na+      Na+  K'
 15                   30                45                 60


http://www.ncbi.nlm.nih.gov/pubmed/?term=18199766
1.25 KH PO , 1 MgSO , 2 CaCl , 16 NaHCO ,
 24                       4                    2                               3

http://www.ncbi.nlm.nih.gov/pubmed/?term=23161880
iological conditions (using HCO  -containing buffers), little or
                                                  3