Modifying session variable breaks get_trackdata()

[kirbyj]

I imagine there is some good reason for this, but I don't understand this behaviour:

> create_emuRdemoData(dir = tempdir())
> demo_data_dir = file.path(tempdir(), "emuR_demoData")
> tg_col_dir = file.path(demo_data_dir, "TextGrid_collection")
> path2directory = file.path(tempdir(), "my-first_emuDB")
> convert_TextGridCollection(dir = tg_col_dir,
                            dbName = "my-first",
                            targetDir = tempdir(),
                            tierNames = c("Word", "Syllable", "Phoneme", "Phonetic"))
> db_handle = load_emuDB(path2directory, verbose = FALSE)

Now create a seglist:

> sl_vowels = query(db_handle, "Phonetic == @")
> sl_vowels
> sl_vowels
# A tibble: 28 × 16
   labels start   end db_uuid  session bundle start_item_id end_item_id level attribute start_item_seq_… end_item_seq_idx type  sample_start sample_end
   <chr>  <dbl> <dbl> <chr>    <chr>   <chr>          <int>       <int> <chr> <chr>                <int>            <int> <chr>        <int>      <int>
 1 @      1506. 1548. af022ed… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124      30968
 2 @      1715. 1741. af022ed… 0000    msajc…           107         107 Phon… Phonetic                22               22 SEGM…        34309      34828
 3 @      1967. 2034. af022ed… 0000    msajc…           112         112 Phon… Phonetic                27               27 SEGM…        39334      40673
 4 @      2303. 2362. af022ed… 0000    msajc…           117         117 Phon… Phonetic                32               32 SEGM…        46059      47238
 5 @      2447. 2506. af022ed… 0000    msajc…           119         119 Phon… Phonetic                34               34 SEGM…        48949      50125
 6 @      1917. 1958. af022ed… 0000    msajc…           118         118 Phon… Phonetic                26               26 SEGM…        38340      39155
 7 @      2022. 2078. af022ed… 0000    msajc…           120         120 Phon… Phonetic                28               28 SEGM…        40439      41569
 8 @      2382. 2431. af022ed… 0000    msajc…           126         126 Phon… Phonetic                34               34 SEGM…        47650      48619
 9 @       330.  380. af022ed… 0000    msajc…            91          91 Phon… Phonetic                 3                3 SEGM…         6609       7590
10 @      1472. 1490. af022ed… 0000    msajc…           108         108 Phon… Phonetic                20               20 SEGM…        29441      29808
# … with 18 more rows, and 1 more variable: sample_rate <int>

Get trackdata:

> td_vowels = get_trackdata(db_handle,
                              seglist = sl_vowels,
                              onTheFlyFunctionName = "forest",
                              verbose = F)
> td_vowels
# A tibble: 287 × 24
   sl_rowIdx labels start   end db_uuid   session bundle start_item_id end_item_id level attribute start_item_seq_… end_item_seq_idx type  sample_start
       <int> <chr>  <dbl> <dbl> <chr>     <chr>   <chr>          <int>       <int> <chr> <chr>                <int>            <int> <chr>        <int>
 1         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 2         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 3         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 4         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 5         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 6         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 7         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 8         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
 9         1 @      1506. 1548. af022edb… 0000    msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124
10         2 @      1715. 1741. af022edb… 0000    msajc…           107         107 Phon… Phonetic                22               22 SEGM…        34309
# … with 277 more rows, and 9 more variables: sample_end <int>, sample_rate <int>, times_orig <dbl>, times_rel <dbl>, times_norm <dbl>, T1 <int>,
#   T2 <int>, T3 <int>, T4 <int>

So far, so good. Now I want session to have some other value:

sl_vowels$session<-recode(sl_vowels$session, `0000`="F1")

# A tibble: 28 × 16
   labels start   end db_uuid  session bundle start_item_id end_item_id level attribute start_item_seq_… end_item_seq_idx type  sample_start sample_end
   <chr>  <dbl> <dbl> <chr>    <chr>   <chr>          <int>       <int> <chr> <chr>                <int>            <int> <chr>        <int>      <int>
 1 @      1506. 1548. af022ed… F1      msajc…           103         103 Phon… Phonetic                18               18 SEGM…        30124      30968
 2 @      1715. 1741. af022ed… F1      msajc…           107         107 Phon… Phonetic                22               22 SEGM…        34309      34828
 3 @      1967. 2034. af022ed… F1      msajc…           112         112 Phon… Phonetic                27               27 SEGM…        39334      40673
 4 @      2303. 2362. af022ed… F1      msajc…           117         117 Phon… Phonetic                32               32 SEGM…        46059      47238
 5 @      2447. 2506. af022ed… F1      msajc…           119         119 Phon… Phonetic                34               34 SEGM…        48949      50125
 6 @      1917. 1958. af022ed… F1      msajc…           118         118 Phon… Phonetic                26               26 SEGM…        38340      39155
 7 @      2022. 2078. af022ed… F1      msajc…           120         120 Phon… Phonetic                28               28 SEGM…        40439      41569
 8 @      2382. 2431. af022ed… F1      msajc…           126         126 Phon… Phonetic                34               34 SEGM…        47650      48619
 9 @       330.  380. af022ed… F1      msajc…            91          91 Phon… Phonetic                 3                3 SEGM…         6609       7590
10 @      1472. 1490. af022ed… F1      msajc…           108         108 Phon… Phonetic                20               20 SEGM…        29441      29808
# … with 18 more rows, and 1 more variable: sample_rate <int>

Note that session is still a <chr>. But now:

> td_vowels = get_trackdata(db_handle,
+                              seglist = sl_vowels,
+                              onTheFlyFunctionName = "forest",
+                              verbose = F)
Error in get_trackdata(db_handle, seglist = sl_vowels, onTheFlyFunctionName = "forest",  : 
  Following utts entry not found: 
In addition: Warning messages:
1: In get_trackdata(db_handle, seglist = sl_vowels, onTheFlyFunctionName = "forest",  :
  The emusegs/emuRsegs object passed in refers to bundles with in-homogeneous sampling rates in their audio files! Here is a list of all refered to bundles incl. their sampling rate: 
[1] session        bundle         media_file     sample_rate    md5_annot_json
<0 rows> (or 0-length row.names)
2: Unknown or uninitialised column: `utts`. 

I am willing to accept that part of the solution here is "don't mess with session, but at the very least, this seems like the wrong error message: nothing has touched the audio files or their sample rates.

[klausj]

I would consider database query results (or at least their session and bundle names) as immutable objects. If you change the session name in the query result, the database will not be updated automagically (by renaming the session). The results are just tibble objects. With the second call to get_trackdata you query bundles of a non-existent session 'F1' and that fails.
I agree that the error message is misleading.
Do you just want to rename the session from '0000' to 'F1' ?

[kirbyj]

We were just using session as a stand-in for "speaker"; it's simpler to just create a new column. But yes, the idea was just to rename the session(s).