17 research outputs found
Figure 3a from: Malavin S, Shmakova L, Claverie J-M, Rivkina E (2020) Frozen Zoo: a collection of permafrost samples containing viable protists and their viruses. Biodiversity Data Journal 8: e51586. https://doi.org/10.3897/BDJ.8.e51586
No full textFigure 3a Late Pleistocene fossil objects on yedoma outcrops sampled for viable protist strains. - Buried Late Pleistocene soil (paleosol) at the Duvannyy Yar outcrop (Sakha Rep.)
Figure 4 from: Malavin S, Shmakova L, Claverie J-M, Rivkina E (2020) Frozen Zoo: a collection of permafrost samples containing viable protists and their viruses. Biodiversity Data Journal 8: e51586. https://doi.org/10.3897/BDJ.8.e51586
No full textFigure 4 Location of sampling events. North-Eastern Eurasia
Frozen Zoo: a collection of permafrost samples containing viable protists and their viruses
No full textInternational audiencePermafrost, frozen ground cemented with ice, occupies about a quarter of the Earth’s hard surface and reaches up to 1000 metres depth. Due to constant subzero temperatures, permafrost represents a unique record of past epochs, whenever it comes to accumulated methane, oxygen isotope ratio or stored mummies of animals. Permafrost is also a unique environment where cryptobiotic stages of different microorganisms are trapped and stored alive for up to hundreds of thousands of years. Several protist strains and two giant protist viruses isolated from permafrost cores have been already describe
kbenoit/quanteda: CRAN v0.9.9-22
No full text<p>Minor fixes in C++ release to comply with CRAN checks on lesser-used platforms.</p>
kbenoit/quanteda: CRAN release v0.9.9-3
No full text<p>Major new update published on CRAN on 2016-01-10. This is a pre-v1.0 release that implements major API changes while still retaining nearly all of the old functions, but hidden and deprecated. See <a href="https://github.com/kbenoit/quanteda/blob/master/NEWS.md">NEWS.md</a>.</p>
kbenoit/quanteda: CRAN release v0.99
No full textNew features
<ul>
<li>Improvements and consoldiation of methods for detecting multi-word expressions, now active only through <code>textstat_collocations()</code>, which computes only the <code>lambda</code> method for now, but does so accurately and efficiently. (#753, #803). This function is still under development and likely to change further.</li>
<li>Added new <code>quanteda_options</code> that affect the maximum documents and features displayed by the dfm print method (#756).</li>
<li><code>ngram</code> formation is now significantly faster, including with skips (skipgrams).</li>
<li>Improvements to <code>topfeatures()</code>:<ul>
<li>now accepts a <code>groups</code> argument that can be used to generate lists of top (or bottom) features in a group of texts, including by document (#336).</li>
<li>new argument <code>scheme</code> that takes the default of (frequency) <code>"count"</code> but also a new <code>"docfreq"</code> value (#408).</li>
</ul>
</li>
<li>New wrapper <code>phrase()</code> converts whitespace-separated multi-word patterns into a list of patterns. This affects the feature/pattern matching in <code>tokens/dfm_select/remove</code>, <code>tokens_compound</code>, <code>tokens/dfm_lookup</code>, and <code>kwic</code>. <code>phrase()</code> and the associated changes also make the behaviour of using character vectors, lists of characters, dictionaries, and collocation objects for pattern matches far more consistent. (See #820, #787, #740, #837, #836, #838)</li>
<li><code>corpus.Corpus()</code> for creating a corpus from a <strong>tm</strong> Corpus now works with more complex objects that include document-level variables, such as data from the <strong>manifestoR</strong> package (#849).</li>
<li>New plot function <code>textplot_keyness()</code> plots term "keyness", the association of words with contrasting classes as measured by <code>textstat_keyness()</code>.</li>
<li>Added corpus constructor for corpus objects (#690).</li>
<li>Added dictionary constructor for dictionary objects (#690).</li>
<li>Added a tokens constructor for tokens objects (#690), including updates to <code>tokens()</code> that improve the consistency and efficiency of the tokenization.</li>
<li>Added new <code>quanteda_options()</code>: <code>language_stemmer</code> and <code>language_stopwords</code>, now used for default in <code>*_wordstem</code> functions and <code>stopwords()</code> for defaults, respectively. Also uses this option in <code>dfm()</code> when <code>stem = TRUE</code>, rather than hard-wiring in the "english" stemmer (#386).</li>
<li>Added a new function <code>textstat_frequency()</code> to compile feature frequencies, possibly by groups. (#825)</li>
<li>Added <code>nomatch</code> option to <code>tokens_lookup()</code> and <code>dfm_lookup()</code>, to provide tokens or feature counts for categories not matched to any dictionary key. (#496)</li>
</ul>
Behaviour changes
<ul>
<li>The functions <code>sequences()</code> and <code>collocations()</code> have been removed and replaced by <code>textstat_collocations()</code>.</li>
<li>(Finally) we added "will" to the list of English stopwords (#818).</li>
<li><code>dfm</code> objects with one or both dimensions haveing zero length, and empty <code>kwic</code> objects now display more appropriately in their print methods (per #811).</li>
<li>Pattern matches are now implemented more consistently across functions. In functions such as <code>*_select</code>, <code>*_remove</code>, <code>tokens_compound</code>, <code>features</code> has been replaced by <code>pattern</code>, and in <code>kwic</code>, <code>keywords</code> has been replaced by <code>pattern</code>. These all behave consistently with respect to <code>pattern</code>, which now has a unified single help page and parameter description.(#839) See also above new features related to <code>phrase()</code>.</li>
<li>We have improved the performance of the C++ routines that handle many of the <code>tokens_*</code> functions using hashed tokens, making some of them 10x faster (#853).</li>
<li>Upgrades to the <code>dfm_group()</code> function now allow "empty" documents to be created using the <code>fill = TRUE</code> option, for making documents conform to a selection (similar to how <code>dfm_select()</code> works for features, when supplied a dfm as the pattern argument). The <code>groups</code> argument now behaves consistently across the functions where it is used. (#854)</li>
<li><code>dictionary()</code> now requires its main argument to be a list, not a series of elements that can be used to build a list.</li>
<li>Some changes to the behaviour of <code>tokens()</code> have improved the behaviour of <code>remove_hyphens = FALSE</code>, which now behaves more correctly regardless of the setting of <code>remove_punct</code> (#887).</li>
<li>Improved <code>cbind.dfm()</code> function allows cbinding vectors, matrixes, and (recyclable) scalars to dfm objects.</li>
</ul>
Bug fixes and stability enhancements
<ul>
<li>For the underlying methods behind <code>textstat_collocations()</code>, we corrected the word matching, and lambda and z calculation methods, which were slightly incorrect before. We also removed the chi2, G2, and pmi statistics, because these were incorrectly calculated for size > 2. </li>
<li>LIWC-formatted dictionary import now robust to assignment to term assignment to missing categories.</li>
<li><code>textmodel_NB(x, y, distribution = "Bernoulli")</code> was previously inactive even when this option was set. It has now been fully implemented and tested (#776, #780).</li>
<li>Separators including rare spacing characters are now handled more robustly by the <code>remove_separators</code> argument in <code>tokens()</code>. See #796.</li>
<li>Improved memory usage when computing <code>ntoken()</code> and <code>ntype()</code>. (#795)</li>
<li>Improvements to <code>quanteda_options()</code> now does not throw an error when <strong>quanteda</strong> functions are called directly without attaching the package. In addition, <strong>quanteda</strong> options can be set now in .Rprofile and will not be overwritten when the options initialization takes place when attaching the package.</li>
<li>Fixed a bug in <code>textstat_readability()</code> that wrongly computed the number of words with fewer than 3 syllables in a text; this affected the <code>FOG.NRI</code> and the <code>Linsear.Write</code> measures only.</li>
<li>Fixed mistakes in the computation of two docfreq schemes: <code>"logave"</code> and <code>"inverseprob"</code>.</li>
<li>Fixed a bug in the handling of multi-thread options where the settings using <code>quanteda_options()</code> did not actually set the number of threads. In addition, we fixed a bug causing threading to be turned off on macOS (due to a check for a gcc version that is not used for compiling the macOS binaries) prevented multi-threading from being used at all on that platform.</li>
<li>Fixed a bug causing failure when functions that use <code>quanteda_options()</code> are called without the namespace or package being attached or loaded (#864).</li>
<li>Fixed a bug in overloading the View method that caused all named objects in the RStudio/Source pane to be named "x". (#893) </li>
</ul>
quanteda/quanteda: CRAN v1.3.0
No full textNew Features
<ul>
<li>Added <code>to = "tripletlist"</code> output type for <code>convert()</code>, to convert a dfm into a simple triplet list. (#1321) </li>
<li>Added <code>tokens_tortl()</code> and <code>char_tortl()</code> to add markers for right-to-left language tokens and character objects. (#1322)</li>
</ul>
Behaviour changes
<ul>
<li>Improved <code>corpus.kwic()</code> by adding new arguments <code>split_context</code> and <code>extract_keyword</code>.</li>
<li><code>dfm_remove(x, selection = anydfm)</code> is now equivalent to <code>dfm_remove(x, selection = featnames(anydfm))</code>. (#1320)</li>
<li>Improved consistency of <code>predict.textmodel_nb()</code> returns, and added <code>type =</code> argument. (#1329)</li>
</ul>
Bug fixes
<ul>
<li>Fixed a bug in <code>textmodel_affinity()</code> that caused failure when the input dfm had been compiled with <code>tolower = FALSE</code>. (#1338)</li>
<li>Fixed a bug affecting <code>tokens_lookup()</code> and <code>dfm_lookup()</code> when <code>nomatch</code> is used. (#1347)</li>
<li>Fixed a problem whereby NA texts created a "document" (or tokens) containing <code>"NA"</code> (#1372)</li>
</ul>
quanteda/quanteda: CRAN v1.4.1
No full textquanteda 1.4.1
Bug fixes and stability enhancements
<ul>
<li>Fixed an issue with special handling of whitespace variants that caused a test to fail when running Ubuntu 18.10 system with libicu-dev version 63.1 (#1604).</li>
<li>Fixed the operation of <code>docvars<-.corpus()</code> in a way that solves #1603 (reassignment of docvar names).</li>
</ul>
quanteda/quanteda: CRAN v1.5.0
No full textNew features
<ul>
<li>Add <code>flatten</code> and <code>levels</code> arguments to <code>as.list.dictionary2()</code> to enable more flexible conversion of dictionary objects. (#1661)</li>
<li>In <code>corpus_sample()</code>, the <code>size</code> now works with the <code>by</code> argument, to control the size of units sampled from each group.</li>
<li>Improvements to <code>textstat_dist()</code> and <code>textstat_simil()</code>, see below.</li>
<li>Long tokens are not discarded automatically in the call to <code>tokens()</code>. (#1713)</li>
</ul>
Behaviour changes
<ul>
<li><code>textstat_dist()</code> and <code>textstat_simil()</code> now return sparse symmetric matrix objects using classes from the <strong>Matrix</strong> package. This replaces the former structure based on the <code>dist</code> class. Computation of these classes is now also based on the fast implementation in the <strong>proxyC</strong> package. When computing similarities, the new <code>min_simil</code> argument allows a user to ignore certain values below a specified similarity threshold. A new coercion method <code>as.data.frame.textstat_simildist()</code> now exists for converting these returns into a data.frame of pairwise comparisons. Existing methods such as <code>as.matrix()</code>, <code>as.dist()</code>, and <code>as.list()</code> work as they did before.</li>
<li>We have removed the "faith", "chi-squared", and "kullback" methods from <code>textstat_dist()</code> and <code>textstat_simil()</code> because these were either not symmetric or not invariant to document or feature ordering. Finally, the <code>selection</code> argument has been deprecated in favour of a new <code>y</code> argument. </li>
<li><code>textstat_readability()</code> now defaults to <code>measure = "Flesch"</code> if no measure is supplied. This makes it consistent with <code>textstat_lexdiv()</code> that also takes a default measure ("TTR") if none is supplied. (#1715)</li>
<li>The default values for <code>max_nchar</code> and <code>min_nchar</code> in <code>tokens_select()</code> are now NULL, meaning they are not applied if the user does not supply values. Fixes #1713.</li>
</ul>
Bug fixes and stability enhancements
<ul>
<li><code>kwic.corpus()</code> and <code>kwic.tokens()</code> behaviour now aligned, meaning that dictionaries are correctly faceted by key instead of by value. (#1684)</li>
<li>Improved formatting of <code>tokens()</code> verbose output. (#1683)</li>
<li>Subsetting and printing of subsetted kwic objects is more robust. (#1665)</li>
<li>The "Bormuth" and "DRP" measures are now fixed for <code>textstat_readability()</code>. (#1701)</li>
</ul>
quanteda/quanteda: CRAN v4.0.2
No full text<p>Minor fixes:</p>
<ul>
<li><p>A failing test caused by C++ code related to <code>fcm()</code> and how tokens objects are re-indexed.</p>
</li>
<li><p>An undeclared package 'quanteda.textstats' in Rd xrefs.</p>
</li>
</ul>
